U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content







AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window






Chapter II



Wounds of the Heart (Including Retained Foreign Bodies), Mediterranean (Formerly North African) Theater of Operations

Lyman A. Brewer III, M.D., and Thomas H. Burford, M.D.


In World War I, as Makins (1) pointed out, casualties with cardiac wounds who survived to come under the care of a surgeon were hit either by relatively small missiles or by missiles traveling with reduced degrees of velocity. The same situation prevailed in World War II. In the 2 years that elapsed between the first Allied landings in North Africa on 8 November 1942 and the end of hostilities in Italy on 2 May 1945, only 75 of the 2,267 thoracic and thoracoabdominal wounds encountered by the teams of the 2d Auxiliary Surgical Group were instances of cardiac or pericardial injury. This is an incidence of 3.3 percent. Of the 75 injuries, 18 were examples of pure pericardial trauma. In the other 57 cases, the heart was involved (table 5). No single team encountered more than 10 cases.

In 1 of the 75 cases, the only stab wound in the series, the wound was self-inflicted. The remainder of the injuries were all battle-incurred, in 53 instances from shell fragments and in 21 from small arms fire. In 2 of the 18 pure peri-

TABLE 5.-Distribution of injuries and anatomic involvements in 56 combat-incurred cardiac injuries1

Type of injury


Anatomic involvement































Contusions and lacerations









Perforating and penetrating









Embolic to heart


















1This table does not include a self-inflicted stab wound, from which the patient recovered. It also does not include 18 combat-incurred pericardial injuries, 3 of which were fatal.


FIGURE 11.-Fatal cardiac contusion. A. Specimen without extensive pericardial damage. B. Cross section showing endocardial thrombi. (a). In this case, rib fragments acted as secondary missiles (b).

cardial injuries, the damage was caused directly by rib fragments, and in at least 1 other case, extensive contusion of the myocardium was caused by rib fragments which acted as secondary missiles (fig. 11). In 43 cases, the injuries were confined to the chest; in the other 32 cases, the wounds were thoracoabdominal.

As these statistics indicate-75 cardiac wounds in 2,267 thoracic and thoracoabdominal injuries-wounds of the heart were not observed with any great frequency in forward hospitals of the Mediterranean theater. The reason, as already intimated, is that they were usually-though not always (fig. 12)-promptly fatal. These figures take no account of immediately fatal wounds, nor do any reliable statistics exist concerning them. In the analysis of 1,000 battlefield deaths by Capt. William W. Tribby, MC (2), it did not prove practical to perform routine autopsies, and even this remarkable series therefore contributes nothing really definite concerning the number of casualties with cardiac wounds who died on the field.

It should be remembered in reading this chapter that, as will be discussed in more detail later, in every instance in this series, the cardiac injury was only a single feature of the trauma. In addition to their cardiac injuries, all of these casualties had sustained more or less extensive wounds of adjacent thoracic structures, and the chest wounds in many instances were further complicated by serious wounds of other parts of the body.


FIGURE 12.-Survivor of cardiac wound shortly after machinegun slug had been removed from his heart. He went on to complete recovery.

The data in these cases, most of which were cared for under stress of battle conditions in forward hospitals, are remarkably complete, thanks to the special records kept by the teams of the 2d Auxiliary Surgical Group. The preservation of essential data, including roentgenograms and electrocardiograms, was facilitated by the peculiar and often dramatic nature of the injury, which attracted the attention of shock officers and internists as well as surgeons. Personal notes on the records, and personal comments by the medical officers who cared for the patients, provided more information in these cases than was usually obtainable from even specially kept clinical records.


Diagnostic Considerations

The preoperative diagnosis of injuries of the heart was not only difficult in itself but was sometimes not made because examination was not directed toward their discovery. One reason for the omission was the impression that


casualties with cardiac wounds did not live long enough to arrive at a forward hospital. As a result, symptoms and signs arising from the cardiac wound, particularly anoxia, were often attributed to other injuries that were present and that complicated the diagnosis. There is no doubt that more careful clinical examinations would have led to an increase in correct diagnosis. It is significant that the same three medical officers made all seven observations of arrhythmia in the 2d Auxiliary Surgical Group series. As with anoxia, the relation of tachycardia to the cardiac wound had to be based on ruling out all other causes for this sign. No single symptom or sign was usually sufficient to establish the diagnosis of cardiac injury.

The patient was usually designated for surgery on the suspicion of a thoracoabdominal wound, or the cardiac wound was found in the course of traumatic thoracotomy or during debridement of a large sucking wound, or it was searched for because of continued intrathoracic bleeding during the course of presumably adequate resuscitation. In 15 of the 75 cases in the 2d Auxiliary Surgical Group series, the injury was discovered at post mortem. In a few cases, it was found, in retrospect, that the symptoms and signs recorded on the chart should have aroused suspicion before operation or autopsy.

Symptoms and Signs

Symptoms and signs were divided into two groups, symptoms due to anoxia and signs suggestive of cardiac dysfunction.

In this series, six patients were recorded as dyspneic, six as needing continuous oxygen, five as mentally confused or semistuporous, and three as cyanotic. Before these findings could be attributed to a cardiac wound, however, other causes of oxygen deficiency, such as hemorrhage, hemothorax, compression pneumothorax, and extensive peritoneal contamination, had to be ruled out. When these conditions had been eliminated and symptoms due to anoxia persisted out of all proportion to visible thoracic damage, then there was justification for regarding them as due to a cardiac lesion.

Signs suggestive of cardiac dysfunction included persistent tachycardia (a pulse of 120 or above) in eight cases, arrhythmia in seven cases (transient fibrillation in one and extrasystoles in six), bradycardia (a pulse below 65) in two cases, an apical systolic murmur and a precordial friction rub in two cases each, a paradoxical pulse in one case, and nausea and vomiting in one case.

Both precordial friction rubs were heard for the first time 24 hours after injury; operation was delayed 3 days in one case and 5 days in the other. Since a friction rub or splash was noted in eight additional cases after operation, it was concluded that a certain length of time must elapse after wounding before this sign appears. The explanation of this phenomenon is the presence in the pericardium of air and fluid. It was seldom seen because, in most instances, the pericardial fluid had drained into the pleural cavity before surgery.


The single instance of paradoxical pulse occurred in a patient with severe myocardial contusion, who died in the shock ward. The persistent nausea and vomiting noted in one case was uncommon in thoracic wounds and led to the suspicion of a thoracoabdominal wound. Whether or not it was caused by the cardiac wound, it was associated with it and made for further diagnostic difficulties.

A precordial crunch or click was occasionally heard, synchronous with the heartbeat. It was associated with mediastinal emphysema and was not considered related to the cardiac trauma.

In retrospect, the most important diagnostic findings were considered to be:

1. A sustained pulse of 120 or more after restoration of satisfactory arterial tension by adequate resuscitation.

2. Continued cyanosis after recovery from shock.

3. Dyspnea out of all proportion to the evident pulmonary pathology. In several patients with cardiac wounds in this series, oxygen had been necessary in the clearing station because of severe dyspnea and cyanosis.

4. The necessity for the early and continuous use of oxygen. These clinical findings were, of course, in addition to:

a. The obvious presence of a precordial wound.

b. The projection of the course of a missile which might reasonably have involved the heart.

In a number of instances, irregularities of cardiac rhythm were not apparent; abnormal cardiac sounds were not heard; and there was a notable absence of significant symptoms and signs, or those present were so trivial as to be misleading. Makins (1) had called attention to these observations in his account of cardiac injuries in World War I. Generally speaking, the safest plan was to suspect cardiac injury whenever a patient seemed generally washed out; had a sustained, rapid pulse with transient irregularities of rhythm; and had a persisting need for oxygen.

Adjunct Diagnostic Measures

Roentgenologic and fluoroscopic studies.-In 15 cases, there was roentgenologic evidence of the injury. In eight cases the films showed a foreign body in the region of the heart, and in four instances the object was termed fuzzy or double-contoured. In five cases the cardiac shadow was altered in size or shape. In two cases the object was thought to be in the region of the heart, but there was no definite proof. In localizing missiles within the cardiac shadow, it was necessary to use heavy penetration (fig. 13), accomplished by the bone technique or the use of the Potter-Bucky diaphragm. A missile could be completely overlooked if the chest was examined with the usual exposure. The interpretation of roentgenologic enlargement of the heart was also open to some question when roentgenograms were made shortly after injury, since there is usually no appreciable stretching of the pericardium when fluid first appears in it.


FIGURE 13.-Foreign body in left ventricle of heart. A. Posteroanterior roentgenogram, with massive left-sided hemothorax obscuring foreign body. B. Same, taken with deep penetration technique, showing foreign body in situ. C. Lateral roentgenogram.

Some cardiac injuries were not suspected until repeated X-ray studies in a general hospital, combined with fluoroscopy, revealed their presence.

The cardiac injury also might be suspected by plotting the probable course of the missile, using the external wounds and fractured ribs as landmarks. Diagnosis was accomplished, or suspicion was aroused, by use of this technique in 22 of the 75 cases in this series. On the other hand, it was easy to be misled concerning a cardiac wound if the foreign body happened to lie free in the pleural cavity.


FIGURE 13.-Continued. D. Fragment after removal.

The diagnostic results might have been better if fluoroscopy had been employed more frequently, since this method permitted observation of possible movement of the object and also made it possible to determine whether it was included in the cardiac shadow in all projections. The cardiac outline was variously described as fuzzy, blurred, enlarged, or of water-bottle shape. In two cases in which the outline was described as blurred or fuzzy, operation revealed hemorrhage within the pericardial membrane and in the areolar tissues of the lower mediastinum.


FIGURE 14.-Foreign body in wall of left ventricle near cardiac apex. A. Posteroanterior roentgenogram. B. Right anterior oblique roentgenogram. The electrocardiogram was consistent with myocardial injury.

Electrocardiography-Neither electrocardiograms nor orthodiagrams were available in forward hospitals. Electrocardiograms were most helpful in determining whether a penetrating wound of the chest had involved the heart (fig. 14), particularly when the thorax was not explored at all or was incompletely explored at debridement. In numerous instances, these studies provided assurance that the myocardium had not suffered injury. In a few cases, progressive alterations in originally normal tracings indicated the necessity for removal of retained foreign bodies.

The most striking abnormalities, as reported by Lt. Col. Edward F. Bland, MC, were observed in myocardial injuries and involved the T-waves and S-T intervals. The predominating pattern was the so-called anterior-apical type, with inversion of T1, T2, and T4 (fig. 15). In the posterior basal type of injury, electrocardiograms were secured in only two cases; in both, there was inversion of T2 and T3. Electrocardiograms with inverted T-waves in all four leads were obtained in five patients. In one of these cases, roentgenograms showed foreign bodies in both the anterior and the posterior cardiac walls. There was no explanation in the other cases for the total T-wave type of inversion.

High-grade conduction defects were not observed in any instance. Two patients with abnormal W-shaped QRS complexes of low voltage and inverted QRS complexes in lead four were thought to have sustained cardiac contusions. In one case, a persistent P-R interval of borderline significance (0.2 second) may have been normal for this particular patient.

In some cases, in which there was roentgenologic proof of the presence of the foreign body in the heart (figs. 16 and 17), electrocardiograms were completely normal.


FIGURE 15.-Serial electrocardiograms taken after penetrating wound of heart. Note the anterior-apical pattern of T1-, T2-, and T4-inversion encountered most often in cardiac wounds.

FIGURE 16.-Machinegun bullet in right side of heart near junction of right auricle and right ventricle. A. Posteroanterior roentgenogram. B. Left anterior oblique roentgenogram. The electrocardiogram showed no abnormalities.


FIGURE 17.-Shell fragment in anterior wall of heart in region of right ventricle. A. Right lateral roentgenogram. The electrocardiogram showed no abnormalities. B. Right anterior oblique roentgenogram.


Hemorrhage into the restricted confines of the pericardial sac, with resulting cardiac tamponade, was the chief danger from a penetrating cardiac wound (fig. 18). Tamponade was, however, relatively infrequent in combat-incurred injuries, in contrast to its frequency in civilian cardiac injuries. The explanation is clear: In civilian life, small weapons, such as knives, icepicks, and small-caliber bullets, are generally used. As a result, the wound is small, bleeding is slow, and tamponade can develop as a physical possibility. In combat-incurred injuries, the wounds are large because the missiles are large, and an outlet for drainage into the pleural cavity is therefore provided. In a few cases, pneumopericardium occurred (fig. 19), without tamponade.

Hemorrhage into the pericardial sac was usually from the cardiac chambers but might also come from a severed branch of the coronary artery, from a pericardial vessel, or from the myocardium itself. The thick muscular wall of the ventricles seemed to tolerate severe lacerations, and even total penetration, without serious bleeding if the coronary arteries, especially arteries with sizable arterial branches, were not involved. In contrast, even slight tears of the thin-walled auricles were apt to be followed by tamponade, which could develop within the space of a few minutes and could be rapidly fatal if it was not promptly corrected. Most casualties in whom hemorrhage was sufficiently rapid to produce early tamponade probably did not live long enough to reach a hospital.


FIGURE 18.-Schematic showing of pathologic physiology of acute pericardial tamponade: Collapse of superior vena cava (a), collapse of pulmonary veins (b), collapse of inferior vena cava (c), impairment of diastolic filling of left and right ventricles (d), impairment of diastolic filling of stria (e), and increase of pressure in jugular vein (f). With these findings, the heart is silent and the pulse pressure decreased.


The diagnosis of cardiac tamponade was generally based on the following findings:

1. Lowered arterial pressure.-In Elkin's (3) series, this finding was present in all cases, and in 17 of the 23, the blood pressure could not be recorded. In all the patients who recovered, the pressure rose immediately after release of the tamponade.

2. Increased venous pressure readings.-These readings, as Elkin (3) noted, are of both diagnostic and prognostic value. If the venous pressure is


FIGURE 19.-Pneumopericardium with retained foreign body just behind heart. A. Posteroanterior roentgenogram, shortly after injury, showing massive pneumopericardium. B. Left (slight) anterior oblique roentgenogram showing retained foreign body. C. Posteroanterior roentgenogram 10 days later, after subsidence of pneumopericardium.

high, the assumption is that the cardiac output is at least sufficient to sustain life. If the venous pressure is low, or falling, the assumption is that the heart is failing and the cardiac output is correspondingly reduced. Three of Elkin's patients who presented this phenomenon died on the operating table or immediately after operation.

3. A quiet heart.-This finding was first described by Bigger (4) and, if it is studied fluoroscopically, the demonstration of decreased cardiac pulsations will be found to be extremely useful in diagnosis.

4. Engorgement of the cervical veins.-This phenomenon, which was present in a number of cases of tamponade, could readily be explained as the result of the inability of these veins to empty into the right auricle which was com-


pressed by the tamponade. When this finding was present, it was pathognomic, since such other causes as congestive heart failure, pulmonary embolism, and mediastinal emphysema were unlikely to be encountered under combat conditions. Absence of distended cervical veins, however, and even of a distended pericardial sac, did not necessarily exclude continued serious hemorrhage from a wounded myocardium. In one case in this series, in which there was continued, vigorous hemorrhage from a myocardial laceration, operation performed 10 hours after injury revealed that tamponade had been prevented by the passage of blood through the pericardial rent into the pleural cavity.

There were 5 instances of tamponade in the 57 cases in which the injuries involved the heart proper. Death occurred in two because the condition was not recognized or suspected. These cases and another case in which survival occurred have features of special interest:

Case Histories

Case l.-This patient was injured in the left chest, shoulder, and buttock by shell fragments on 14 December 1944. He was not in evident shock and at a field hospital, his condition was listed as good. Breath sounds were diminished over the left chest. He was given 1,000 cc. of physiologic salt solution intravenously before operation, which was performed 7 hours after injury. All wounds were debrided. An open pneumothorax was closed without exploration. Aspiration of a hemothorax yielded 300 cc. of blood.

The patient's postoperative condition was good. The pulse was within normal range, the blood pressure was 90/60 mm. Hg, and there was full recovery from the anesthesia. Four hours later, he was turned in bed, at his own request. Shortly afterwards, he was found dead. Post mortem showed the pericardium to be distended with 200 cc. of blood and the heart to be constricted. A perforation of the superior portion of the pericardium on the left side, about 1.5 cm in diameter, was occluded by a fibrinous exudate. A metallic fragment 1.5 by 1 by 1 cm. was found in the pericardial cavity. In the right ventricle was a laceration 1.5 by 1 cm in length and 3 to 4 mm. in depth. A small branch of the right coronary artery was severed. Two small mural thrombi were found in the right ventricular cavity beneath the laceration. The liver was moderately congested. The cause of death was obviously cardiac tamponade, which had not been suspected before the post mortem examination.

Comment-The operating surgeon, who witnessed the post mortem, expressed the opinion that this soldier's life could have been saved if (1) thoracotomy with control of the coronary bleeding had been carried out, or (2) if there had been closer postoperative observation for signs of developing tamponade, which could have been relieved by aspiration (fig. 20). It is possible that if tamponade had been relieved, hemorrhage might have ceased spontaneously. More likely, exposure and ligation of the severed artery would have been necessary.

Case 2-This patient was in poor condition when he was received in a field hospital after sustaining a bullet wound of the right lower quadrant of the abdomen. The neck veins were swollen and the heart sounds were barely audible. Although cardiac tamponade was suspected, no blood could be aspirated from the pericardium, and roentgenograms failed to disclose a foreign body in or near the heart. At operation, 15 hours after injury, a tear in the colon was sutured, and a gutter wound of the liver was drained. Left thoracotomy was then performed. The pericardium was distended with clotted blood. The heart was constricted, and its enfeebled action soon ceased, despite vigorous efforts at


FIGURE 20.-Management of cardiac tamponade by aspiration: Substernal transdiaphragmatic aspiration (a), and left lateral aspiration (b).

resuscitation. Autopsy, which was performed immediately, revealed a .30-caliber bullet lying in the right ventricular cavity, surrounded by a foul-smelling clot. The bullet had been deflected upward from the pelvis and had passed through the colon and the liver before perforating the diaphragm and the right ventricular wall.

Comment.-A combination of unfortunate circumstances partly explained the fatal issue in this case. The negative results of two ordinarily reliable precautionary measures served to disarm suspicion. One was the pericardial tap, which was dry because the blood was clotted. The other was roentgenologic failure to demonstrate the missile within the thorax, probably because it was obscured by cardiac motion. Under these circumstances, the swollen cervical veins became extremely important from the diagnostic standpoint. They always indicated serious circulatory imbalance which demanded correction before other surgical procedures could be undertaken with safety. Whether or not this patient's life could have been saved by initial thoracotomy is debatable, but the fact that he survived 18 hours after wounding and withstood the added strain imposed by an extensive laparotomy before the chest was explored is highly suggestive. Had the operative procedures been reversed, it might have been possible to relieve tamponade, check myocardial hemorrhage, and perhaps remove the intraventricular bullet. Even if its removal had proved impossible, it might have migrated later to the pulmonary artery without necessarily serious consequences, as in another case in this series (case 9, p. 82).

Case 3.-This patient sustained a penetrating shell-fragment wound of the left lumbar region, with a fracture of the tenth rib, on 8 July 1944. He was received at the field hospital in moderate shock, with a blood pressure of 90/50 mm. Hg. At operation, 8 hours after injury, a segment of the ninth rib was resected, and 1,400 cc. of blood was evacuated from the pleural cavity. A hole in the pericardium 2.5 centimeters in diameter was extended to reveal a laceration of the same extent in the left ventricle, from which a small stream


of blood exuded with each heartbeat. The hemorrhage was controlled and the laceration closed by two figure-of-eight silk sutures in the ventricular wall. The pericardium was left open for a distance of 6 centimeters. The foreign body was found in the lower lobe of the left lung, with a hematoma but no hemorrhage. When the diaphragm was opened between the anterior and the posterior perforations, the spleen was found fractured and actively bleeding. It was removed. Two perforations of the stomach were sutured. The diaphragm and the chest wall were then closed.

The patient's immediate postoperative condition was good, the blood pressure being 110/70 mm. Hg. Ten days later, his condition was still good; he had no complaints, and temperature, pulse, and respiration were normal. There was no fluid in the pleura, but a pericardial friction rub was noted. The following day, the patient was transferred to a general hospital. On 23 July, he suddenly complained of severe dyspnea and precordial pain and became extremely cyanotic. The lungs were filled with rales. The pulse was 130. Gradual improvement occurred following the administration of morphine and oxygen, and 4 hours later, the lungs had cleared. The medical consultant made a diagnosis of acute left ventricular failure.

The following day, an electrocardiogram showed late inversion of T1 and T4, with low voltage of the QRS complexes, consistent with recent left ventricular injury and probably pericarditis. By 25 July, the patient was much improved; the heart sounds were good, and he had no cardiac symptoms. Electrocardiograms on 8 August and on 21 August were still abnormal, but the inversion of the T-waves was less marked. On 11 September, 8 weeks after injury, he was evacuated to the Zone of Interior in good condition.

Comment-This case is of special interest on two counts: (1) In spite of serious injury to the heart, lungs, stomach, and spleen, this patient survived. His convalescence was undoubtedly hastened by early and adequate surgery. (2) This is the only case in the entire series of cardiac injuries in which delayed acute left ventricular failure occurred (on the 15th day). The precise explanation for this isolated and unexpected episode remains obscure. In this case, as in several others, cardiac tamponade was prevented by the escape of blood through a sizable pericardial rent. It is probable that the cardiac hemorrhage accounted for the major portion of the 1,400 cc. of blood evacuated from the left chest at operation.


Cardiac injuries encountered at surgery or autopsy were classified as contusions, lacerations, lacerations and contusions, penetrating wounds of the cardiac chambers with retained foreign bodies, and perforating (through-and-through) wounds. Cardiac emboli were also possibilities. These injuries were additionally classified according to the particular cardiac structure involved.

Superficial abrasions of the epicardium and engorgement or thrombosis of the subepicardial vessels were often found. When a major artery was badly contused, thrombosis was a possibility.

The myocardium might show gross evidence of degeneration or actual necrosis. Microscopically, interstitial hemorrhage varied in variety and extent. The muscle fibers showed fragmentation, loss of striation, or advanced necrosis. Eosinophilia, leukocytic infiltration, and beginning phagocytic removal of necrotic muscle tissue were observed as early as 18 hours after injury.

When the endocardium was injured or there was subendocardial hemorrhage, adherent mural thrombi might develop. This was observed in five cases


in this series, in three instances at autopsy. When extensive lesions were scattered along the acute or the obtuse margins of the heart, the hemorrhage might involve the myocardium of both ventricles and extend into the interventricular septum.

The high rate of energy imparted to the tissues in the track of the missile explained why particles of tissue were thrown laterally and passed their energy on, thus producing further damage. Attention has already been called to the cases in which fragments of the ribs produced this sequence of events, which was especially likely to occur in tangential wounds.

In some instances, the missile passed through the myocardium at such an angle that the cardiac walls closed behind it.


General Considerations

There were 18 instances of pure pericardial injury in this series, with 3 deaths, all of which occurred more than 48 hours after wounding and none of which was due primarily to the pericardial injury. In 14 instances, the wounds were lacerated; and in 5, foreign bodies were present, in 2 instances consisting of rib fragments. One of these missiles was in the free sac.

In one case, injury of the pericardium was associated with an injury of the myocardium; it is quite remarkable that this combination of injuries did not occur more often, as some of the pericardial lacerations were severe. It is also quite possible that minor lacerations of the pericardium occurred and remained undetected. The academic question also arises as to the probable considerable margin of safety afforded by a heart in systole as compared with one in diastole at the moment of injury (Wood (5) and Nicholson's "near misses").


The mere diagnosis of cardiac tamponade was not regarded as an absolute indication for surgical intervention. In several cases, prompt recovery followed pericardial aspiration alone (fig. 20), sometimes after a single aspiration. The general opinion, however, was that it was hazardous to depend routinely upon this type of conservative management, especially if an irregular foreign body with sharp margins or points was demonstrated in either the myocardium or pericardium. Under the circumstances, if more than two satisfactory aspirations were necessary, it was considered that surgery was indicated. The possibility of infection also had to be considered (p. 356).

In urgent cases, management consisted of repeated aspiration of the pericardium until surgery could be undertaken. The needle was inserted in the angle between the xiphoid process and the adjacent left costal arch and was directed cephalad, inward, and toward the left, at an angle of about 45.


Repair was accomplished through a curved incision exposing the third, fourth, and fifth costal cartilages; resection of sufficient portions of these ribs to exposed the pericardium; incision of the pericardium; and suture of the wound by the technique recommended by Elkin (3) and by Beck (6).

In 5 of the 18 pericardial injuries, the pericardium was sutured tightly. In the remaining cases, drainage was instituted into the pleural cavity. In two of these five cases, in one of which there was an associated myocardial wound, there was massive, troublesome pericardial effusion postoperatively, a complication not observed in any case that was drained. Recovery in both cases followed paracentesis. All five foreign bodies were removed. In two other cases, in which it was thought that foreign bodies might be present in the pericardium, no attempt was made to remove them.

In one case, the pericardial sac was enormously distended with blood. It was incised for a distance of 10 cm. from the superior to the inferior margin, and there was no further bleeding after the first gush of blood. The pericardium was left open and the chest was closed. Recovery was uneventful. The origin of the tamponade in this instance was obscure, but the azygos vein, although it is usually extrapericardial and therefore an unusual source of intrapericardial hemorrhage, was considered the most probable source.

Pneumopericardium was observed in three of the pericardial injuries, as the result of the entrance of air into the pericardial sac from a pneumothorax, injured lung, bronchus, or esophagus. In two cases, the air was promptly absorbed, without evidence of pericardial irritation. In the other case, fibrous pericarditis developed.

While a pericardial wound was sometimes relatively innocuous, at least as compared to the lethal potentialities of other cardiac wounds, it could be extremely urgent. This is clear if the pathologic process (fig. 18) be considered: With fluid in the pericardium, an obstruction exists to the filling of the heart, and the blood is dammed up in the great venous channels of the body. If the intrapericardial pressure comes to equal the effective venous pressure, the blood can no longer enter the right auricle and death will occur promptly. That is why so many deaths from this cause must be assumed to have occurred on the battlefield and why, in some instances, the condition might be so urgent that the time required to confirm a suspected diagnosis by roentgenologic or fluoroscopic examination might mean the difference between life and death.


General Considerations

The 16 contusions observed in the 2d Auxiliary Surgical Group series of cardiac wounds were similar to the contusions described by Elkin (3), Beck (6), and others in civilian life, as the result of blunt trauma to the chest, steering wheel injuries, and similar accidents.


There were 11 deaths in the 16 cases, in 6 of which the cardiac state was considered entirely responsible for the fatal outcome.

The diagnosis of a cardiac contusion was not particularly difficult. Most of the patients with significant injuries of this type presented signs and symptoms indicative of oxygen want and cardiac dysfunction, that is, tachycardia and arrhythmia. These signs and symptoms, as well as the gross and microscopic appearance of the myocardium at autopsy, had much in common with the clinical and pathological picture of myocardial infarction following coronary occlusion.

Pathogenesis and Pathologic Process

In most of the battle casualties, the damage resulted from the force propagated by the passage of a small, high-velocity missile in the immediate vicinity of the heart. In at least one instance of myocardial contusion (fig. 11), the ribs were shattered, and their fragments apparently acted as secondary missiles, with resultant direct blunt injury. Thrombus formation followed, and death ensued; the pericardium was intact.

Whether a localized blast effect resulted from the passage of the missiles in these cases cannot be answered. Certainly, in this series, no cardiac contusion resulted from the generalized effect of a pressure wave in the atmosphere. While it was theoretically possible for a serious cardiac injury to result from blast, there was no confirmation at autopsy, in which the possibility was borne in mind in the examination of patients who died of blast injuries.

Clinical and experimental studies before the war had clarified the pathologic physiology of thoracic contusions. Bright and Beck (7), as well as Warburg (8), had demonstrated that trauma to the intact chest, especially in the young adult whose chest is far more resilient and flexible than the chest of an older person, may be directly transmitted to the heart, whether the injury is a direct blow over the precordial region, compression of the chest between two solid objects, or a blow over the abdomen, with a sudden rise in intra-abdominal pressure.

Experimental trauma most often caused immediate rupture of one of the cardiac chambers, with death. Contusion, with resultant petechial hemorrhage, softening and necrosis of tissue, and eventual rupture, might involve any portion of the conduction system. Most frequently, it caused reflex spasm of the coronary vessels, with the production of a syndrome like the common civilian-type syndrome of coronary occlusion and infarction. It was also shown experimentally that in cases of vagosympathetic imbalance, or when the heart was sensitized by adrenalin, cardiac irregularities and coronary spasm resulted more readily. In the heat of battle, when the vascular system was surcharged with considerable adrenalin, slight trauma to the chest might conceivably cause considerable cardiac disturbance.

The pathologic process in contusive lesions consisted of scattered or confluent petechial hemorrhages involving the myocardium over various areas of one or both chambers (fig. 21). In these 16 cases, the ventricles were involved


FIGURE 21.-Scattered petechial and confluent hemorrhages of right ventricle caused by the indirect force of a penetrating bullet wound of sternum.

15 times (7 left, 5 right, 3 both) and the right auricle, once. Superficial abrasions of the epicardium and engorgement of the subepicardial vessels were sometimes observed. The myocardial hemorrhage often extended through to the endocardium, and the muscles sometimes showed gross evidence of degenerative changes or actual necrosis. In fatal cases, in which there had been involvement of the entire thickness of the myocardium, mural thrombi were frequently found attached to the endocardium (fig. 11). When extensive lesions were scattered along either the acute or the obtuse cardiac margin, it was not uncommon to find hemorrhages extending into the myocardium of both ventricles, and even into part of the interventricular septum. The pericardium was not necessarily injured in myocardial contusion; it was intact in 9 of the 16 cases in this series. The pathologic pattern in the fatal cases was essentially one of subpericardial and subendocardial hemorrhage, usually petechial in distribution.


A patient with a myocardial contusion was a poor risk for any kind of surgery, especially during the first several hours after wounding. The fact that 6 of the 11 deaths in these 16 cases were due to the cardiac lesion indicates just how poor risks these casualties were.


The contusion itself was not a surgical lesion, but, unfortunately, numerous associated wounds often required that surgery be done promptly. Ideally, it was postponed for 24 to 48 hours, but this was frequently impossible. Certain wounds, such as thoracoabdominal wounds, required prompt operation, in spite of the fact that immediately after injury, the risk of death from irritability of the myocardium and potentially lethal arrhythmias might be entranced by anesthesia or surgical manipulations. The best that could be done was to delay surgery as long as possible, to permit some degree of cardiac recovery.

In the meantime, the patient was treated as if he were suffering from acute coronary occlusion. Resuscitation was carried out, and surgery was then undertaken with due realization that the risk was inevitably great and the mortality would be correspondingly high. In purely thoracic wounds, in which surgery was not mandatory within 6 to 12 hours, it was best to delay it as long as possible.

In three fatal cases in this series, in which operation was performed, respectively, at 5, 11, and 17 hours after wounding, it was thought that further delay might have been beneficial. Two of the wounds were purely thoracic. The other was a high thoracoabdominal wound, in which it was clear that only the liver was involved. In each instance, signs of cardiac dysfunction were prominent. During resuscitation, the patients were in poor general condition, semistuporous, with rapid pulse, and dyspneic out of all proportion to the visible intrathoracic damage. In each instance, the systolic pressure was elevated to 95 mm. Hg or higher, but death occurred on the operating table or immediately after surgery was concluded.

In another case of this kind, not included in this series, the timelag was only 6 hours. Death occurred on the operating table. At autopsy, it was confirmed that the wound was purely thoracic. There was extensive contusion of the right ventricle and thrombosis of the anterior descending branch of the left coronary artery.

Any of these patients might well have died, even if surgery had not been done, but the added burden of the anesthetic and the operative procedures cannot be ignored in assessing the outcome in seriously wounded patients.

In two other cases in this series, which may be cited in contrast, surgery was delayed for 3 and 5 days, respectively, after wounding. Both patients were received in shock. In one case, the pulse remained over 120 for 48 hours. In the other, for 4 days, there were intermittent periods of cardiac arrhythmia, associated with wet lung, pulmonary edema, and jaundice. In each instance, the surgeon expressed the opinion that the patient might well have died if operation had been done even as late as 12 hours after wounding.

The final conclusion was that if a cardiac contusion were diagnosed and such indications for early surgery as continuing hemorrhage or a thoracoabdominal wound did not exist, surgery should be deferred for at least 24 to 48 hours, to provide every opportunity for the reduction of myocardial irritability.


FIGURE 22.-Electrocardiogram of patient with cardiac contusion showing W-shaped QRS complexes of low voltate and inverted QRS-4. The same electrocardiographic phenomena were observed in another casualty with a cardiac contusion.

Electrocardiograms in two survivors of contused wounds were identical with respect to (1) abnormal W-shaped QRS complexes of low voltage in the limb leads and (2) inverted QRS-4 (fig. 22). This pattern was not observed in any other type of cardiac injury. These electrocardiograms also showed the abnormalities of T-waves and S-T intervals frequently seen after myocardial injury.


General Considerations

Of the 20 lacerations in this series, 10 were combined with contusions of the myocardium and 4 were detected only at post mortem. In the 10 pure lacerations, the left ventricle was affected in 7 cases, the right ventricle in 2, and the right auricle in 1. These 10 cases, in 2 of which there were foreign bodies in the myocardium, comprise all the instances in the series of incised, cleanly lacerated wounds of the myocardium in which there was no gross evidence of myocardial contusion or necrosis.

In the combined lacerations and contusions, the left ventricle was affected five times, the right ventricle twice, and both ventricles twice. In the remain-


FIGURE 23.-Technique of cardiac surgery. A. Laceration of right ventricular wall without serious contusion. B. Direct suture of laceration of ventricular wall with digital control of bleeding.

ing case, the right auricle and right ventricle were injured. The serious implications of contusions are again evident in this group, in which there were five deaths, four due to the cardiac wound, against the single death due to a cardiac lesion in the pure lacerations.


An analysis of the records shows that a rather surprising number-10-of the 16 lacerations of the myocardium were not repaired, and apparently with no immediate ill effects. In the four lacerations found only at post mortem, there was no evidence in any instance that the fatality was due to failure to effect a repair.

In the six cases which were repaired, the laceration was completely closed in four by suture. In the two other cases, complete approximation was impossible, and the pericardium was used to help bridge the defect; free muscle grafts were also used (figs. 23 and 24).


FIGURE 24.-Technique of cardiac surgery. Closure of penetrating wound of heart. A. Wound of right bentricle. B. Use of free muscle graft (greatly enlarged). C. Cross section of sutured muscle graft.

The pericardium was either sutured over the wound or sutured to the edges of the poorly approximated wound. In the former instance, posterior drainage was employed.

The number of cases is too small to permit drawing any conclusions as to the wisdom of operating in forward hospitals solely to suture cardiac lacerations.


General Considerations

There were 19 instances of perforating or penetrating wounds of the heart in this series, 7 involving the left ventricle, 3 the right ventricle, 2 the left auricle, and 7 the right auricle. There were nine deaths, in eight instances due to the cardiac lesion. In one wound of the auricle, discovered only at post mortem, death occurred 24 hours after wounding. In the opinion of those who witnessed the autopsy, the fatal outcome was due to extensive wounds elsewhere in the body. It was thought that the patient would have survived without auricular repair. The most frequent complication, and the most important cause of death, was hemorrhage, which could readily be exsanguinating.


It is of great interest that through-and-through cardiac wounds were not always immediately fatal. In one case in this series (p. 75), recovery followed the repair of two wounds of the left ventricle caused by a shell fragment that divided the left phrenic nerve and the pericardiophrenic vessels, entered the left ventricle at the apex, and made its exit on the posterior wall. Operation was performed 14 hours after wounding.

One other patient with a through-and-through wound of the left ventricle also survived. The three perforations of the right auricle all ended fatally. In the experience of the 2d Auxiliary Surgical Group, no patient with a perforation of the interauricular or interventricular septum survived to reach a forward hospital.


The chief indication for immediate surgery in wounds of the cardiac chambers was continuing hemorrhage. If bleeding caused tamponade rather than exsanguinating hemorrhage, as it sometimes did, treatment could be more individualized. If the tamponade developed rapidly, it was considered better to operate at once, particularly if it was known that the missile causing the wound was large. If the tamponade developed slowly, one or two aspirations might be attempted (p. 64).

If foreign bodies were encountered in the course of the operation, an attempt was made to remove them. If their presence was suspected, an attempt was made to locate them. If, however, they were not found immediately, the correct procedure was to control the hemorrhage and close the chest. Since hemorrhage was the indication for operation, a long-continued search for the missile, with blind manipulations within the cardiac chambers, could not be considered justified. The foreign body could be removed later at the base hospital if that proved necessary.

Complete closure of the cardiac wound was possible in 10 cases. One wound had ceased to bleed when it was exposed, and suture was not considered necessary. In two instances, both auricular wounds, attempts at closure failed, and both patients died of intractable hemorrhage. Attempts to plug the defect with the finger were unsuccessful.

Case Histories

Case 4.-This 22-year-old private was wounded by an artillery shell fragment on 27 June 1944. At the field hospital, 2 hours later, he was found to be mildly shocked, but he was conscious and in fairly good condition. Examination revealed a lacerated wound, 3 cm. long, in the fifth left intercostal space, just outside the midclavicular line. There was no dyspnea or hemoptysis. The heart tones were normal, and no adventitious sounds were heard. There were signs of fluid in the left pleural cavity. The abdomen was tender and resistant to pressure.

The patient became nauseated and vomited twice shortly after admission. A Levin tube was passed into the stomach, and 250 cc. of air and fluid were withdrawn; there was no blood in the gastric contents.


FIGURE 25 (case 4).-Repair of through-and through wound of heart. A. Posteroanterior roentgenogram showing large left hemothorax, which obscures lung. Shell fragment at level of ninth rib posteriorly measured 23 by 10 by 4 mm. (insert). A Levin tube can be seen faintly at the level of the eleventh intercostal space of the left. B. Drawing depicting anterolateral laceration of pericardium with severance of pericardiophrenic artery and phrenic nerve, as follows: Incision for exposure (a), incision for flap (b), wound of entrance (c), and wound of exit (d). Note relation of wounds to important blood vessels.


FIGURE 25.-Continued. C. Closure of anterior heart wound. Apex of heart held by right hand of assistant, exposing anterior laceration. D. Reinforcement of anterior cardiac wound by suture of pedicled graft of pericardium.

After 500 cc. of plasma and 1,000 cc. of blood had been given over a 2-hour period, the blood pressure rose from 80/60 to 130/70 mm. Hg, and the pulse fell from 110 to 90. Abdominal signs and symptoms persisted. Roentgenologic examination (fig. 25A) showed the shell fragment lying well posterior, apparently just within the costal cage. From the location of the wound, the apparent course of the missile, and the persisting abdominal signs, especially the nausea and vomiting, a thoracoabdominal wound could not be ruled out.

Operation was carried out under endotracheal anesthesia 6 hours after wounding and 3 hours after hospitalization. Blood transfusion was continued during the procedure.


FIGURE 25.-Continued. E. Posteroanterior roentgenogram 8 weeks after operation, showing clear lung fields, no cardiac enlargement, and elevation and paralysis of left diaphragm. Note chip fracture of fifth rib anteriorly (arrow).

After debridement, anterior thoracotomy was performed in the fifth intercostal space by extension of the wound of entrance. A chip fracture of the fifth rib was found, with a small contusion of the lingula of the left upper lobe, but the diaphragm was intact. Five hundred cubic centimeters of blood were evacuated from the pleural cavity, after which the shell fragment was discovered lying free posteriorly. Two perforations were seen in the pericardium. The pericardiophrenic artery and phrenic nerve had been severed anterolaterally (fig. 25B), and a posterior laceration was found just lateral to the reflection of the parietal pleura from the pericardium onto the mediastinum. The pericardial sac was opened by vertical incision a few minutes after 5 cc. of procaine hydrochloride had been injected into it. Two lacerated wounds of the left ventricle, each 8 mm. in length, were found oozing blood with each systole. The wound of entrance was at the apex and the wound of exit in the midportion of the left ventricular wall posteriorly (fig. 25B).

The apex of the heart was rotated 90 forward and steadied by the right hand of the assistant (fig. 25C). The wound was exposed between the spread of his second and third fingers, and bleeding was controlled by the application of two silk sutures (No. 0). A small venous branch was occluded by the sutures.

The anterior laceration was irregular, and the muscle gaped slightly. Oozing continued, particularly after a suture had cut partly through the muscle. It was controlled by suturing a small pedicled graft of anterior pericardium and pericardial fat over the laceration (fig. 25D).

Extrasystoles were numerous while the heart was manipulated but ceased immediately when manipulations were discontinued. Crystalline sulfanilamide and penicillin were placed in the pleural cavity, and drainage was instituted by means of two water-trap tubes. A small mushroom catheter was used in the second intercostal space anteriorly


and a quarter-inch fenestrated tube in the eighth intercostal space in the midaxillary line. The tube was clamped off for 8 hours after operation, to permit contact of the chemotherapeutic agents with the tissues. The incision was closed in layers, without pericostal sutures.

At the conclusion of the operation, the blood pressure was 110/70 mm. Hg. the pulse 145, and the respiration 32. Standard postoperative measures were employed. Recovery was generally smooth. The blood pressure remained normal; the pulse stabilized at 100-110. Cardiac irregularities were never evident.

Twenty-four hours after operation, a splash, synchronous with systole, was heard over the precordium. Four days later, a loud precordial friction rub was audible for 24 hours; during the same time, the second sound in the pulmonic area was occasionally reduplicated. On the sixth day, there was a slight roughening of the first sound at the apex, and a poorly transmitted, soft systolic murmur was heard in this area.

When the patient was transferred to a general hospital on the 12th postoperative day, he was afebrile, with a pulse of 88 and a blood pressure of 115/60 mm. Hg. He became ambulatory in another week. A scratching to-and-fro precordial friction rub was heard intermittently for another 2 weeks.

Roentgenologic examination 8 weeks after operation showed the lung fields clear and the heart normal in size (fig. 25E). Electrocardiograms 2 weeks after operation showed moderate inversion of T-waves with elevated S-T intervals in the first three leads and moderate left axis deviation (fig. 26). Two weeks later there was less inversion of T1. Another electrocardiogram 2 weeks after the first showed T1 upright, but T2 and T3 remained deeply inverted, and left axis deviation persisted.

No signs of cardiac embarrassment developed as the patient increased his activities, and he was in excellent condition when he was evacuated to the Zone of Interior 11 weeks after injury.

Comment-This case is remarkable because, in spite of the through-and-through wound of the heart, the patient survived, after early, adequate surgery. Convalescence was quite smooth. It is noteworthy that in spite of the double wound in the ventricle, the blood loss was minimal. It is also noteworthy that no signs of cardiac weakness ensued in spite of the multiple wounds elsewhere, plus the necessary manipulation of the heart at operation and the repeated intravenous infusions of blood and saline solution that were necessary.

The serial electrocardiograms available in this case were thought to be the first on record taken during recovery from a complete perforation of the heart. It is of some interest that the effect of the T-waves in leads II and III of the posterior (basal) injury appears to have overshadowed the effect of the apical wounds (leads I and II), possibly because of the greater mass of muscle injured in the penetration of the thicker basal wall of the ventricle.

That this patient survived his original wound seems due to a happy and unusual chain of circumstances. The fragment must have struck end-on and passed through the heart without revolving. It also seems probable that the perforation occurred during diastole, so that the chamber was traversed by the missile without irreparable damage to the papillary muscle.

As this case also demonstrates, there may be few if any localizing signs or symptoms in spite of a serious cardiac wound. The nausea and vomiting were probably cardiac in origin; these symptoms do not usually occur in pure thoracic injuries. With the shell fragment free in the pleural cavity, the course of the missile was misleading, and the wrong conclusions concerning it were drawn. Had the missile come directly from the front, it could not have failed to penetrate the diaphragm. It came, however, from the left, passed through the heart, and then fell free in the pleural cavity, almost opposite the wound of entrance in an anteroposterior plane.


FIGURE 26 (case 4).-Repair of through-and through wound of heart. Serial electrocardiograms 12 days, 17 days, 41 days, and 56 days after wounding. Note that residual inversion of T2 and T3 overshadows the temporary inversion of T1 (T2) pattern of anteriorapical injury.


Case 5-In another ventricular injury, the patient died in the shock ward of a field hospital 6 hours after he had received a penetrating wound of the left chest, before operation could be performed. Autopsy revealed that a single shell fragment, 1 by 1 by 1 cm., had entered the chest through the left scapula, fractured the third and fourth ribs posteriorly, and then perforated the wall of the left ventricle, to become embedded in the opposite ventricular wall. Death was caused by hemorrhage from the heart into the left chest. The rent in the pericardium had permitted the escape of blood and thus prevented tamponade.

Comment-The comment of the medical officer who performed the autopsy was that this man had lived for 6 hours with a hole in his heart and that his life might have been saved if adequate blood had been available and if the chest had been opened at once.


General Considerations

The intravascular migration of projectiles and other foreign bodies to the heart and pulmonary circulation from distant wounds by way of the great veins is so uncommon as to constitute a true medical curiosity.1 Although still rare, these migratory objects were recognized more frequently in World War II than in the past. The need for their removal was still the subject of some disagreement. Some objects remained asymptomatic for long periods of time, but others, because they served as foci of infection or caused damage to the myocardium, caused death from embolism. Early removal in a base section center was considered the wisest plan if cardiac disability or other clinical signs and symptoms were present. Otherwise, the policy was to return the patients to the Zone of Interior.

If the fragment entered the pulmonary circulation from the right heart, it was theoretically possible for it to serve immediately as a fatal embolus. Removal of a missile from the pulmonary vessels could be attended with great difficulty and might require the sacrifice of an essential artery, perhaps with lobectomy or pneumonectomy. In one case in the 2d Auxiliary Surgical Group experience, surgery was not carried out for this reason (case 9, p. 82).

Migratory foreign bodies were not necessarily fatal. Three of five patients observed in the Mediterranean theater are known to have recovered, and at least one of the two deaths was not caused by the presence of the foreign body, while the same comment is possibly applicable to the other case. The case histories follow.

Case Histories

Case 6-A soldier who sustained a penetrating shell-fragment wound of the right cervical region on 17 June 1944 was treated conservatively. Roentgenograms revealed a

1Although only one instance of migratory foreign body seems to have been reported in World War I (9), undoubtedly others occurred. In the reported case, a shell fragment buried in the liver was left in situ at operation. The patient died of peritonitis 6 days later. At necropsy, the metal fragment was found covered with fibrin and enmeshed in the columnae carneae. Examination of the liver disclosed the track of the fragment into its point of entry into a large hepatic vein and its subsequent intravascular passage to the heart.


foreign body in the right mid lung field, with no evidence of pulmonary injury. He was discharged to duty on 21 July but on 9 August was readmitted to the hospital, complaining of dyspnea and a vague pain in the right chest on effort. Roentgenograms showed the foreign body in the same position as at the first examination. At exploratory thoracotomy on 18 August, the missile was found in the right inferior branch of the pulmonary artery. It was palpated through the wall of the artery but was not removed. On 10 October, when the patient was evacuated to the Zone of Interior, he was ambulatory but was still mildly dyspneic, and he continued to complain of vague chest pains on effort.

Comment.-When this patient was first admitted to the hospital, it was thought that the missile had entered the thorax directly from the cervical wound. In the light of the findings at operation, however, it must be assumed that it penetrated the right subclavian vein, migrated to the right side of the heart, and thence passed to the right pulmonary artery. No infarction resulted, and it was considered unlikely that it would cause serious trouble at its present site.2

Case 7-This man sustained penetrating wounds of the right flank on 11 October 1944, with a resulting perforation of the colon and another of the left chest, with hemothorax.

FIGURE 27 (case 7).-Migratory intravascular foreign body. Lower portion of inferior vena cava showing wound of entry and superimposed metal fragment that had passed as embolus to branch of left pulmonary artery.

2As already pointed out (p. 17 fn), postwar experimental studies by Dr. Lyman A. Brewer III and his associates showed that metallic foreign bodies in the lobar branches of the pulmonary artery are usually well tolerated, provided that the bronchial arteries remain intact. Those occluding the main pulmonary arteries cause serious pulmonary changes and should be removed.


The following day, laparotomy and colostomy were performed, and the chest wound was closed. On 25 October, he died from an intraperitoneal hemorrhage. Necropsy revealed multiple penetrating wounds, including a puncture wound, 0.8 cm. in diameter and surrounded by necrotic tissue, on the posterior wall of the inferior vena cava just above its formation. An embolic shell fragment 1.5 by 0.6 by 0.4 cm. was found in the inferior branch of the left pulmonary artery, with no evidence that it had entered through the lung parenchyma (fig. 27). There was no associated pulmonary infarction. The anatomic cause of death was massive hemorrhage secondary to the abdominal wound.

Comment-The metallic embolus to the lung found in this case at autopsy was a complete surprise. There had been no signs or symptoms suggestive of its presence before death, and roentgenologic study had not seemed justified because of the patient's continued precarious state. It is entirely probable that he would have recovered from the vascular injury had he not succumbed, 2 weeks after wounding, to the secondary complications of his severe abdominal injuries.

Case 8-This patient sustained penetrating wounds of the right shoulder and the right lumbar region on 2 October 1944. Though laparotomy revealed that a foreign body had entered the abdomen, it could not be found. Subsequent roentgenograms suggested that it was on the right side and lying in the psoas muscle. On 7 October, the roentgenograms were repeated because the patient had continued to run fever and had developed rales on the left side. The film was hazy, and the foreign body seen in the first film on the right side now appeared to be above the diaphragm on the left side.

Two days later, the temperature rose to 104 F. Other findings included a rapid pulse (from 140 to 160); a blood pressure of 95/50 mm. Hg; a variable systolic sound of unusual character over the heart to the left of the lower sternum, somewhat suggestive of a friction rub; and pulmonary rales. There was no venous distention. Roentgenograms revealed the foreign body apparently lying anteriorly and inferiorly in the pericardium and slightly blurred by motion. On 12 October, the patient continued to have fever, pulmonary rales were still present, and the superficial systolic "noise" persisted. His condition deteriorated progressively, and he died before a definite diagnosis was made, though the pyrexia was thought to be due to malaria.

At autopsy, when a penetrating wound in the right flank was explored, a track was found which indicated that the foreign body had entered the inferior vena cava. Continued exploration revealed it lying free in the right ventricle (fig. 28A). The entrance of the missile into the inferior vena cava was represented by an oval defect 1.5 by 1.5 by 5 cm. in the posterior wall above the bifurcation (fig. 28B). The defect was surrounded by numerous thrombi, some of which appeared to have become detached. The fragment found in the right ventricle near the apex was large and irregularly oblong. It measured 2.3 by 1 cm. and weighed 10.35 grams. There was slight discoloration of the adjacent endocardium, but no thrombi were found. Careful examination of the heart and vena cava revealed no evidence of penetration, and there was no doubt that the metal fragment had entered the vascular system through the wound in the lower vena cava. Both lungs showed widespread areas of infarction, and multiple emboli were demonstrable in the pulmonary arterial branches. In addition, there was massive intraperitoneal hemorrhage (1,500 cc.) from an omental vessel.

Comment-In spite of the size and weight of this jagged fragment, it had apparently churned about in the right ventricle for over a week without causing serious injury to the cardiac wall. Although it was apparently too large to pass through the pulmonary orifice, at no time did the patient exhibit acute seizures suggesting a ball-valve effect, and at no time were the cervical veins distended. The persistent fever and downhill course were adequately explained by the recurrent pulmonary emboli from the mural thrombi in the lower vena cava, but these phenomena were probably unrelated to the presence of the foreign body in the right ventricular cavity.


FIGURE 28 (case 8).-Migratory intravascular foreign body. A. Large oblong shell fragment free in chamber of right ventricle. B. Lower portion of inferior vena cava showing wound of entry surrounded by thrombi just above bifurcation. Fragment in ventricle is shown above.


FIGURE 29 (case 9).-Migratory intravascular foreign body. A. Posteroanterior roentgenogram showing shell fragment in left hilar region. B. Left lateral roentgenogram localizing foreign body to intrahilar pulmonary area. The foreign body could not be found in this position or elsewhere at operation.

Case 9-When this soldier was wounded in action on 12 April 1944, he sustained multiple severe penetrating wounds of the right thorax, right leg, and both feet. Initial roentgenologic study showed a large metallic foreign body in the left lung. The wounds were debrided.

On 28 April, when he was transferred to a thoracic surgery center, his condition was good except for moderate dyspnea, which continued after cardiorespiratory disequilibrium had been treated by adequate thoracentesis and nerve block. Localization studies showed the fragment to be in the hilar region of the left lung (fig. 29A and 29B). Fluoroscopy on 9 May showed it to be in the root area on this side.

At thoracotomy the following day, the foreign body could not be found, and the chest was finally closed after a long and fruitless search for it. Roentgenograms taken immediately after operation revealed it lying in the right hilar region (fig. 29C).


FIGURE 29.-Continued. C. Posteroanterior roentgenogram immediately after left thoracotomy showing foreign body in right hilar region. D. Right lateral roentgenogram showing fragment again localized in root area of lung. At operation immediately afterward, it was found in the lumen of the main right pulmonary artery, where it was left in situ.

After operation, the patient remained more dyspneic than seemed warranted by the findings in the chest. The only change noted on electrocardiography was a moderate sinus tachycardia.

On 9 July, a right-sided thoracotomy was done, after repeated roentgenograms and fluoroscopy immediately before operation had shown the fragment still in the right hilar region (fig. 29D). The missile was found impacted within the lumen of the main right pulmonary artery. A palpable thrill was felt over it and for a short distance into the artery distally. The pulmonary circulation was entirely adequate. Complete dissection of the hilar structures did not achieve sufficient mobilization of the artery to warrant an


attempt to remove the missile, since the involved segment of the artery lay directly beneath the superior pulmonary vein anteriorly and upon the right stem bronchus posteriorly. Since there was no evidence of aneurysmal dilatation of the artery or of inadequacy of the pulmonary circulation, it was decided not to sacrifice the posterior pulmonary vein in order to remove the foreign body.

Convalescence was uneventful except for a disproportionate degree of dyspnea for a few days following operation. The patient was transferred to the Zone of Interior on 16 August, ambulant and in good condition. In October, a followup letter from the United States reported that he had continued well, with no symptoms other than dyspnea when he walked rapidly. No further roentgenologic studies and no operative procedure had been carried out.

Comment-This case might be fairly termed unique in medical annals. The opportunity for adequate roentgenologic study and for complete exploration of both hilar regions by a competent chest surgeon left no doubt that the large metal fragment had migrated, against the blood flow in the pulmonary circuit, from its original position in the left pulmonary artery to its subsequent lodgment in the right pulmonary artery, without causing serious symptoms or recognizable complications. The exact mechanism of the migration is difficult to explain. It seems unlikely that it was accomplished by gravity alone. It may have been the result of manipulation during the exploration of the left hilar region, but no definite statement can be made on this point.

It is a matter of special interest that during a review of this case shortly after the first thoracotomy, the operating surgeon stated that the missile had probably entered a major lobar radicle of the right pulmonary artery and had progressed from there.3

Case 10-This patient sustained a machinegun bullet wound which penetrated the right chest, just below the middle third of the clavicle, on 31 May 1944. On 4 June, he was transferred from a field hospital to a general hospital. He was then in fair condition except that his temperature was 102 F., the heart sounds were distant, and the veins of the neck were distended. Roentgenograms showed an enlargement of the heart shadow and a foreign body 1.7 by 1 cm. inside the left cardiac border. In the right chest was a moderate collection of fluid. On 7 June 1944, there was considerably increased distention of the neck veins, and the venous pressure was 210 mm. H2O. Electrocardiograms showed QRS complexes of low voltage. Pericardicentesis produced 750 cc. of old, bloody fluid. The patient showed prompt improvement but had a pericardial friction rub for the next 3 days. Roentgenograms showed a smaller heart shadow, together with a surprising shift of the foreign body, which now lay just inside the right border of the heart. By 20 June, the patient was ambulatory. His improvement was progressive and satisfactory except that he complained of a dull pain running from the sternum to the cardiac apex when he walked. Roentgenograms showed the foreign body now lying behind the cardiac apex just inside the left border. On 2 August, the foreign body, which was embedded in dense pericardial adhesions behind the apex, was removed, with considerable difficulty. On 2 October, the patient was evacuated to the Zone of Interior fully ambulatory and complaining only of slight substernal ache on effort.


The single stab wound in the series, which was self-inflicted, was similar to the same type of injury observed in civilian practice. Emergency surgery, 16 hours after injury, was required to control hemorrhage from a small punc-

3This patient was seen again in 1948, at which time he was perfectly well. In April 1959, he suffered profuse hemoptysis because the foreign body had eroded through the right pulmonary artery into the right bronchus intermedius. Pneumonectomy was necessary, from which he made a good recovery. He has remained in good health to date (October 1960).


ture wound of the left ventricle near the left descending coronary artery. A laceration of the myocardium was also sutured. The postoperative electrocardiograms showed inversion of Tl, T2, and T4. The patient was evacuated to the Zone of Interior in good condition 8 weeks after operation.


Since lacerations and penetrations of the great vessels were apt to be quickly fatal, it is not surprising that the experience of the 2d Auxiliary Surgical Group included only two such injuries. The first patient died on the operating table, from cardiac tamponade. The pericardium was distended by hemorrhage from a small laceration of the superior vena cava, and the fragment was recovered free in the pericardial sac.

The second patient was of interest because of the retained foreign body which rested in close apposition to the ascending aorta and moved vigorously with each cardiac pulsation (fig. 30). The vessel wall was presumed to have escaped injury, and the patient was returned to the Zone of Interior without operation overseas.

That this patient was likely to continue well was evident in the case history of another soldier observed in the Mediterranean theater, who had harbored a .38-caliber bullet directly against the ascending aorta for the past 20 years (fig. 31). On fluoroscopy, the bullet was seen to move vigorously with each pulsation of the aorta. In the interim, he had had no symptoms of any kind, and he was returned to the Zone of Interior not because of the presence of the bullet but because of bronchial asthma.


Official Policies

The infrequency of wounds of the heart susceptible to surgical management is implicit in the scant attention paid to them in the instructions for wound management issued during World War II. They are not mentioned in the circular letters published in either the Mediterranean or the European theaters or in the "Manual of Therapy" published in the European theater just before D-day. They are also not mentioned in War Department Technical Bulletin (TB MED) 147, which was published in March 1945 and which dealt with the care of battle casualties in the light of the wartime experience to date.

In fact, the only detailed instructions for management of wounds of the heart appeared in the thoracic surgery section of the military manual on neurosurgery and thoracic surgery published in 1943 under the auspices of the Committee on Surgery of the Division of Medical Sciences of the National Research Council (10).


FIGURE 30.-Retained shell fragment in close apposition to ascending aorta. It moved vigorously with each pulsation. A. Posteroanterior roentgenogram. B. Right anterior oblique roentgenogram. C. Lateral roentgenogram.

Policies in the Mediterranean Theater

Although foreign bodies within the heart were not a major problem in base installations in the North African campaigns or later in the Italian campaigns, it early became necessary to establish a policy concerning their management. Discussions by Col. Edward D. Churchill, MC, Consultant in Surgery to the theater surgeon, with the thoracic surgeons in the theater led to the establishment of the following principles:

1. Only foreign bodies in the heart that were causing significant clinical symptoms or giving rise to significant clinical signs were to be removed in over-


FIGURE 31.-Retained .38-caliber bullet lying directly against ascending aorta, in which position it had been for the past 20 years. Although it moved vigorously with each pulsation of the aorta, the patient had never had symptoms referable to it. A. Posteroanterior roentgenogram. B. Lateral roentgenogram.


sea base installations. All such casualties were to be sent to chest centers, for management by qualified thoracic surgeons.

Two groups of objects were believed likely to give trouble. The first was composed of missiles, particularly low-velocity fragments, lying in the myocardium, without complete penetration of the chambers. These, it was thought, would give rise to continuing intracardiac bleeding or to nonhemorrhagic pericardial effusion or to both. A number of such cases were managed successfully by surgery. The second group was composed of intramyocardial foreign bodies impinging on major coronary vessels, which might give rise to symptoms of coronary arterial insufficiency. Two such cases were treated, with complete relief of anginal symptoms after removal of the objects.

2. It was not believed that so-called chamber foreign bodies; that is, foreign bodies lying in an auricle or ventricle, would be a problem in an oversea theater. On the basis of long-term followup reports it was thought that patients in this group could be safely evacuated to the Zone of Interior for such surgery as might be necessary. On the other hand, if a chamber foreign body migrated intravascularly to the pulmonary vessels, removal in the oversea theater was considered indicated.

Experience confirmed the validity of this policy. Only a small number of true chamber foreign bodies were observed, none of which migrated to the pulmonary vessels or gave rise to difficulty during the period of observation overseas.

Surgical Timing

When a patient with a suspected cardiac wound was encountered in wartime, the decision as to its management could not be based, as in civilian life, solely upon the presence of the wound and the patient's status. In wartime, the heart wound frequently represented only one of several injuries, and both the diagnosis of the cardiac wound and the timing of surgery for it were complicated by the presence of these multiple injuries. As already noted, only half of the cardiac injuries in this series were diagnosed before operation or post mortem, but, in view of the difficult circumstances, the percentage of diagnoses is gratifying rather than otherwise.

Decision as to surgery rested upon two considerations, the type of injury and the presence or absence of a foreign body in the heart or pericardium. Three questions had to be answered:

1. Could the cardiac lesion itself be corrected by surgery?

2. What was the effect of the patient's cardiac status on his ability to withstand surgery for other wounds?

3. Should the surgery be performed at a forward hospital, at the base, or in the Zone of Interior?

When the heart was exposed for any reason at initial wound surgery in a forward hospital, suture of the laceration was best accomplished at the forward hospital. Efficient repair was unlikely in the base section after 10 days


or more had elapsed after the injury. After this lapse, retraction of the edges of the myocardial defect and induration from proliferation of fibroblastic tissue combined to defeat good approximation.

Foreign bodies, however, presented a different problem. When it was certain that a foreign body identified by roentgenologic examination was just within the myocardium or had merely penetrated the pericardium, it was best to postpone surgery for 7 to 14 days and to evacuate the patient to a base center unless there were early and continued episodes of bleeding or cardiac dysfunction. Continued bleeding from a cardiac wound was always an indication for prompt thoracotomy. Otherwise, surgery was seldom an emergency. The dearth of diagnostic facilities and the lack of time for unhurried study in forward installations made accurate localization in them difficult or impossible.

Specifically, indications for delayed removal of foreign bodies included cardiac pain, arrhythmia, abnormalities in previously normal electrocardiographs, and suspected intrathoracic hemorrhage.

Preoperative Management

Resuscitation of patients with recognized cardiac wounds followed standard principles. Originally, in accordance with the work of Beck (6), it had been thought that if tamponade existed or was suspected, it would be of no value to give blood or any other fluid intravenously, since it could not reach the heart. The experimental work of Cooper, Stead, and Warren (11) shortly before the war had shown that rapid intravenous infusion, with the subsequent increase in blood volume, enabled dogs to withstand considerably higher intrapericardial pressure than when this measure was omitted. Elkin (3) had also shown its clinical value. Rapid blood transfusions were therefore strongly recommended as part of the preoperative routine.


When a cardiac wound was considered in need of surgical repair or when such a wound was suspected, adequate exposure through an elective approach was mandatory. The tragedy of inadequate exposure was well illustrated by a case in the 2d Auxiliary Surgical Group experience in which exsanguination occurred from an unsuspected wound of the right auricle. The surgical incision, a low posterior thoracotomy, was intended for a thoracoabdominal wound and was not suitable for control of auricular hemorrhage.

Surgeons of the 2d Auxiliary Surgical Group believed very strongly that extrapleural techniques should not be employed, however desirable they might be in civilian life. In the Duval-Barasty type of extrapleural surgery, both auricles and ventricles were exposed simultaneously, and the argument was advanced that with such exposure, there was less possibility for missiles to


FIGURE 32.-Technique of cardiac surgery. A. Left anterolateral transpleural incision to expose left ventricle. B. Exposure of right ventricle aided by transection of sternum.

migrate from one chamber to the other. Exposure for posterior lesions, however, was not satisfactory by the extrapleural technique. Such operations always took longer than operations performed by the transpleural technique, and there was no time for them in the busy wartime operating room. Finally, the intrapleural damage and hemothorax almost always present in these battle-incurred wounds made extrapleural techniques so difficult that they were entirely impractical.

Surgical Approach

An anterior transpleural approach (fig. 32A) was generally most suitable. An intercostal incision was always employed unless the rib in the involved area was badly fractured. The third or fourth intercostal space provided the best exposure for wounds of the auricle and the fifth or sixth interspace for wounds of the ventricle. In general, more of the right ventricle could be exposed through a left-sided than through a right-sided thoracotomy (fig. 32B).

The incision was carried to the sternum, and the internal mammary vessels were ligated and divided. When it proved necessary, an increase in vertical exposure could be obtained by transverse section of the sternum at the level of the intercostal incision.

Usually, according to the suggestion of Beck (6), 10 cc. of 5 percent procaine hydrochloride was injected into the pericardial sac several minutes before the heart was exposed. This practice materially cut down the incidence of ectopic beats while manipulations were in progress.


As a rule, the pericardium was found tense, with the cardiac pulsations weak or even imperceptible. If the pericardial wound was found, it was enlarged as necessary. If it was not found promptly, the pericardium was opened between stay sutures. Once intrapericardial pressure was relieved, bleeding became more profuse and cardiac contractions increased in force. The heart wound was occasionally located before blood and clots were removed and before the heart began to beat actively. Suture was simple under these conditions. More often, the wound was not located until free blood and clots had been removed by suction.

Control of Hemorrhage and Rotation of Heart

In ventricular wounds, if the left index finger was placed over the wound, bleeding was usually controlled sufficiently to permit the passage of a suture directly under the finger. The suture, which was left untied, was held in the left hand, so that hemostasis could be secured by traction while other sutures were placed and tied. If the wound was on the diaphragmatic surface, on the posterior aspect of the heart, or behind the sternum, a stay suture was sometimes passed through the apex, by Beck's (6) technique, so that the heart could be rotated into a favorable position for suture of the wound.

Surgeons of the 2d Auxiliary Surgical Group found that the hand of the assistant (fig. 25C) made a better retractor than any suture. By this means, the apex of the heart could be rotated forward at least 90, and cardiac movement was considerably dampened. Spreading the fingers provided a sliding type of retractor which permitted exposure of any portion of the cardiac wall.

If the lesions were anterior, the palming method (Sauerbruch grip) had advantages. By this technique, the third, fourth, and fifth fingers of the surgeon's left hand were passed behind the heart. The index finger was anterior, and the thumb was used as necessary for hemostasis. This technique provided excellent control of both the heart and the bleeding area.

Suture Techniques

Much of the wartime knowledge of actual cardiac suture techniques was owed to the prewar work of Beck (6), Elkin (3), and Bigger (4), in particular (fig. 33). The general techniques which they had promulgated pointed to the direction, and furnished the background, of most cardiac surgery.

Some lacerations in this series were deliberately not sutured, for two reasons: Some were considered too slight to require repair, and some were so located, or were of such a character, that it was thought that attempts at suture might lead to additional difficulties. Among the wounds left unsutured were (1) superficial lacerations of the myocardium 1 or 2 mm. in depth, which were not bleeding, particularly if the left ventricle was involved; (2) round or oval lacerations left after removal of foreign bodies, especially in the apical region; and (3) lacerations near a major coronary artery, in which the risk of


FIGURE 33.-Technique of cardiac surgery. A. Closure of perforating wound of left auricle by modified Beck technique. B. Closure by modified Elkin technique. Occlusion of wound with rubber-shod forceps.

nonrepair had to be weighed against the chances of thrombus formation. Wounds of the coronary vessels were not necessarily fatal. If small branches were bleeding, very careful ligature or suture of the individual branches was highly satisfactory.

Wounds which had penetrated the cardiac chambers were always sutured, even though they were plugged by blood clots and were not bleeding when they were exposed. If they were not sutured, secondary hemorrhage was always a possibility.

In general, it was easier to repair lacerations of the auricle on the right side than on the left. Since the right wall is thinner than the left, repair of wounds in this area was always necessary, on the ground that after repair, the wall was thicker, and the scar from the sutured laceration stronger, than a naturally healed wound. Complete suture or some other type of repair was considered mandatory if the depth of the laceration felt thin or if there was any myocardial bulging. Without adequate suture, aneurysm of the myocardium might develop later. Cases of this kind were reported by Loison (cited by Lilienthal (12)).

Large wounds of the auricle deserve special mention. Even wounds as large as 3 cm. might not be exsanguinating because the lung had collapsed against the wound or a clot had formed. The maneuver of covering the defect with the finger in penetrating wounds, as practiced in ventricular wounds, was,


however, not practical in wounds of the auricle because of the thinness of the auricular wall. If sutures could not be placed immediately, the best plan was to grasp each edge of the laceration with fine forceps, which were then approximated, or to use temporary ligatures until sutures could be properly introduced. If the wound was at the edge of the auricle, it would be completely occluded with rubber-shod forceps.

Perforating through-and-through wounds of the chambers could be repaired successfully if both wounds were superficial. No method of exposure was practical for repairing a wound of the posteromesial surface of the right auricle. The only two patients in the 2d Auxiliary Surgical Group series with this type of wound both died of exsanguination, caused by hemorrhage from the free wounds. If the cardiac wound could have been repaired, the mediastinal perforation might have sealed off.

A needle with a small eye or an atraumatic needle was used. Sutures were interrupted and were usually of braided No. 00 or No. 000 silk, preferably oiled or waxed. They were placed as close as possible to the wound edges but did not include them, if this could be avoided (it could not always be avoided in auricular wounds), because of the possibility of thrombus formation. In myocardial defects, sutures were taken into the epicardium and superficial myocardium. No sutures included the endocardium (fig. 23). Great care was taken in the placing of sutures, since necrosis of the wound edges, particularly when the cardiac chambers were involved, could lead to fatal necrosis and secondary hemorrhage. Sutures were tied during systole when possible and always without tension.

Some lacerations, because of loss of substance as the result of the necrosis and contusion of surrounding tissues, were difficult if not impossible to suture completely. Then considerable ingenuity had to be exercised to secure closure, particularly when the laceration extended into a cardiac chamber. Free muscle grafts were useful, and probably should have been employed more frequently than they were. They could be laid in the defect, where they were held in place by fine sutures (figure 24B). They helped to fill the defect and were also instrumental in checking hemorrhage or myocardial oozing.

As a further reinforcing mechanism, the pericardium was sutured over the area of repair, after establishment of drainage for the pericardial sac into the pleural cavity by a cruciate incision. The edges of the pericardium could be either approximated or imbricated (fig. 24C). This technique, combined with the use of a muscle graft, provided solid repair. A flap of pericardium as an extra layer was also useful when bleeding had not been completely controlled by myocardial suture.

Drainage was provided into the pleural cavity by closing the pericardium so loosely that fluid was able to escape. In the two cases in which this precaution was omitted, pericardial effusion occurred (p. 65).

The chest wall was sutured with careful approximation of the anatomic layers.




Lacerations and perforations-Twenty-nine cardiac injuries were visualized at operation and managed as follows:

Complete repair was accomplished in four lacerated wounds (one of the left ventricle, two of the right ventricle, and one of the right auricle and right ventricle).

Partial repair was accomplished in two lacerations of the left ventricle.

Complete repair was accomplished in 10 perforating or penetrating wounds (6 of the left ventricle, 1 of the right ventricle, 1 of the left auricle, and 2 of the right auricle).

Repair was attempted but proved impossible in two perforated wounds of the right auricle.

No repair was attempted in 11 wounds, including 10 lacerations (8 of the left ventricle, 1 of the right ventricle, and 1 of the right auricle), and 1 perforating wound of the right ventricle.

Foreign bodies-It was the policy in the Mediterranean theater to be extremely conservative about the removal of foreign bodies in forward hospitals or, indeed, in any oversea hospital. Of the 29 retained missiles in the 75 cardiac wounds in the 2d Auxiliary Surgical Group experience, only 9 were removed, as follows:

Four (of four) were recovered from the pericardium.

One (of three) was recovered from the pericardial sac.

Three (of ten) were recovered from the myocaridum; five others were found at autopsy.

One (of four) was recovered from the cardiac chambers; three others were found at autopsy.

The majority of fragments not removed were small, 0.5 cm. or less in diameter. In some cases, the condition of the patient did not warrant an extended search for them. The two foreign bodies not removed from the pericardial sac were not identified positively, but from the roentgenologic evidence and operative findings, it was considered highly likely that they were present.

Of the eight foreign bodies found only at autopsy, the missile was directly responsible for one death, because of an embolus to the heart (case 7, p. 80) and possibly for a second death (case 8, p. 80).

Postoperative Observations and Complications

Certain cardiac abnormalities were observed in some of the patients who survived surgery:

Eight developed friction rubs, some of which were audible for as long as 3 weeks.


Two patients had apparent myocardial accidents, presumed to be infarctions. One occurred in the single stab wound in the series (p. 84). The other patient developed a typical coronary occlusion 24 hours after operation, with transient auricular fibrillation, precordial pain, and circulatory collapse. This patient had presented extrasystoles before operation, and at operation, it was necessary to ligate a small bleeding terminal branch of the anterior descending artery. In addition, he had a clean superficial laceration of the cardiac apex, which was not repaired.

One patient had a transient pneumopericardium.

One patient developed hemiplegia after repair of a laceration of the left ventricle. One may speculate whether mural thrombi formed after operation.

Two patients developed massive pericardial effusions. In neither instance was the pericardium drained at operation. Both did well with pericardicentesis, and there were no apparent sequelae.

Four other patients had delayed pericardial effusions between 2 and 6 weeks after injury, for no apparent reason. All recovered under conservative management. The clinical course suggested an irritative phenomenon rather than a flareup of latent infection. British observers in the Mediterranean theater, who observed a few similar cases, accepted this development as an indication for removal of any retained foreign bodies as soon as the acute reaction subsided. The limited U.S. experience, as well as theoretical considerations, suggested a more benign interpretation, which remained for the future to clarify.

Fibrinous pericarditis developed in four cases and quite likely was present in others in which it was not clinically apparent. The removal of dressings and bandages to establish this fact would not have been justified. In the affected patients, electrocardiograms usually exhibited the expected changes in the QRS complexes (low voltage), with lesser degrees of alteration in the T-waves and S-T intervals.

Purulent pericarditis occurred in one case in the series in which the diagnosis of pericardial injury was missed at first:

Case 11-This patient sustained a severe penetrating shell-fragment wound of the epigastric region on 3 June 1944. Debridement revealed no pleural penetration, and no foreign body was visualized. Laparotomy revealed no intra-abdominal injury.

Five days after wounding, the patient became increasingly dyspneic, and cardiac enlargement was noted. Digitalis was administered, on a diagnosis of myocardial insufficiency. When he was admitted to a thoracic surgery center on 11 June, 600 cc. of dark, bloody fluid, without odor, was removed from the pericardium, with prompt relief of dyspnea. Systemic penicillin therapy was begun. Five days later, only 15 cc. of fluid was obtained on pericardial aspiration, but 2 days later, 1,000 cc. was obtained. The patient became progressively more febrile, and on 17 June, the fluid removed (400 cc.) had an offensive odor, showed early purulent transition, and on direct smear was found to be teeming with organisms.

Pericardiostomy was performed immediately, by removal of segments of the fourth and fifth costal cartilages. A large metallic foreign body, surrounded by a large piece of cloth, was found in the posterior recess of the left pericardial sac; it had driven


FIGURE 34 (case 11).-Suppurative pericarditis. A. Posteroanterior roentgenogram showing extent of accumulated fluid in pericardium. Aspirated fluid showed early purulent transition and had a foul odor. B. Same, 3 weeks after pericardial drainage.

itself partly through the pericardium and into the left pleural space. The missile, with its cloth investment, was removed, and the pericardium was emptied by suction and was tacked to the pectoral fascia to keep it widely open. No other provision for drainage was made. Cultures were reported positive for proteolytic clostridia.

The patient made a very rapid recovery, and the pericardial cavity was rapidly obliterated. There were no episodes of recurrent dyspnea or paradoxical pulse. The fever subsided promptly, and the patient was up and about 18 days after operation.

Comment-The foreign body in this case was not visualized (fig. 34A) by roentgenograms because of the superimposition of the heart shadow and the density of the fluid in the pericardium. Its presence in the pericardium was suspected because of the failure to visualize it elsewhere and the nature of the wound. The patient's rapid convalescence is explained by the fact that drainage was instituted early, before the pericardial fluid became thick and intensely purulent. As a result, there was no recurrence of dyspneic episodes and no irregularities of the pulse. There was also prompt obliteration of the pericardial sac (fig. 34B). Fibrinous deposition of exudate on serosal surfaces is an important factor in delayed convalescence in such cases and furnishes an urgent reason for early drainage.


There were 30 deaths in the 75 cardiac injuries managed by the surgeons of the 2d Auxiliary Surgical Group, 3 among the 18 pericardial injuries and 27 among the 57 myocardial injuries.

All three deaths in the pericardial group occurred 48 hours or more after operation, and none could be attributed to the cardiac wound per se. Of the 27 deaths in the myocardial injuries, 20 were, however, directly attributable to the cardiac wounds. The seven remaining fatalities were variously attributable to associated wounds or to shock, bronchopneumonia, and anuria.

Ten cardiac wounds were seen only at autopsy, as follows:


Four lacerated wounds, including one of the left ventricle, one of the right ventricle, and two of both ventricles.

Six perforated wounds, including one of the left ventricle, one of the right ventricle, one of the left auricle, and three of the right auricle.

Causes of death-Exclusive of the 3 deaths in the 18 pericardial injuries, the 27 deaths in the 57 wounds of the heart proper were distributed as follows:

Eleven deaths, six due to cardiac causes, in sixteen pure contusions. Three of these cardiac deaths were due directly to myocardial lesions and usually occurred suddenly, from infarction or arrhythmia. In the three other cases, the heart wound played an essential contributory part in the fatality but was not the only cause.

Five deaths, four due to cardiac causes, in the ten lacerations associated with contusions.

One death, due to the myocardial lesion, in the ten pure lacerating wounds.

Nine deaths, eight due to cardiac causes, in the nineteen perforating or penetrating wounds. In four cases, the fatality was due to exsanguinating hemorrhage from the cardiac wounds, and in two cases to tamponade.

One death, in a myocardial lesion, due to a cardiac embolus.

An analysis of the cardiac deaths (table 6) shows that 15 of the 20 followed wounds of the ventricles.

Time of death-The distribution of the 27 deaths in the series in relation to the time of the fatality was as follows:

Two before operation, both as the result of the cardiac wound.

Ten, including four in thoracoabdominal injuries, before surgery was completed or immediately after operation. Eight were considered due to the cardiac wound.

Four, including three thoracoabdominal wounds, between 1 and 5 hours postoperative. Three were considered due to the cardiac wounds.

TABLE 6.-Distribution of 27 deaths in 56 combat-incurred cardiac injuries1

Type of injury

Number of cases

Number of deaths

Anatomic distribution of cardiac deaths




























Contusions and lacerations









Perforating or penetrating









Embolus of heart


















1This table does not include a self-inflicted stab wound, from which the patient recovered. It also does not include 18 combat-incurred pericardial injuries, 3 of which were fatal.


Two between 6 and 12 hours postoperative. Neither was due directly to the cardiac wound.

Six, including one thoracoabdominal wound, between 12 and 24 hours postoperative. Four were considered due to the cardiac wound.

Three, including two thoracoabdominal wounds, after 24 hours. All were considered due to the cardiac wound.

Of the 20 deaths considered directly due to the cardiac wound, 2 occurred before operation, 8 during or immediately after operation, 3 within 5 hours after operation, 4 within 13 to 24 hours postoperative, and 3 after 24 hours postoperative.

Responsibility for cardiac fatalities-In 20 deaths attributed to wounds of the heart, either surgery was not done at all, or corrective surgery was not completed. An analysis of these cases indicates that in eight, surgery would probably have benefited the patients and that in one other case, it might possibly have been beneficial. In the other 11 cases, surgery would not have been useful, or more extensive surgery than was carried out would have been useless. Of these 11 casualties, 4 were in such poor condition when they were first seen that they died promptly, either before or during operation. Another, thought to be in good condition, died suddenly during operation, and three others, also thought to be in good condition, died suddenly soon after operation. The other three patients in this group died after operation, soon after symptoms of myocardial infarction had become evident.

When the analysis of the possible benefits of surgery, or of more complete surgery, is made in these 20 cases from the standpoint of the causative lesion, the following facts emerge:

1. The six casualties with contusions could not have benefited by operation. This is not a lesion which is ever benefited by surgery, nor is it a condition in which early operation for other causes, however necessary it may be, is well tolerated.

2. Surgery would not have been useful in the single fatal lacerating wound uncomplicated by other cardiac damage. The patient had a pericarditis produced by contamination with gastric contents, though the exact role of the infection in the fatal outcome cannot be stated precisely.

3. Operation would not have been useful in four cases of myocardial lacerations and contusions in which it was omitted, nor would additional surgery have been beneficial in a fifth case in which the laceration was repaired. Damage was lethal in all four of these cases. In three, the anterior descending branch of the left coronary artery presented a traumatic thrombosis for at least half its length, and the fourth patient had an early pericarditis produced by contamination from gastric contents. A fifth patient, with a myocardial laceration and contusion, might possibly have been helped by surgery. He died suddenly, 4 hours after debridement of the thoracic wound. At autopsy, a shell fragment 15 by 10 by 10 mm. was found in contact with the sternum, lying in a shallow, contused, lacerated wound of the right ventricle at the base of the pulmonary


conus. It is possible that the continued presence of this missile was responsible for propagating fatal ectopic stimuli. The pericardium contained 150 cc. of liquid blood, but the sac was not tense, and the mode of death did not suggest that pressure from tamponade played any part in the fatal outcome.

4. Operation would probably have been useful in the fatal case in which the foreign body became embolic to the heart.

5. The largest group of cases in which surgery might have been beneficial consisted of seven penetrating or perforating wounds of the cardiac chambers. Repair of the defect would probably have succeeded in six of these cases and might have helped in the seventh. Four of the patients died of exsanguination, one in the shock ward, two on the operating table, and the other of an unrecognized perforating wound of the right auricle. In this case, original hemorrhage had apparently ceased when debridement of the thoracic wall was done, but rapid exsanguination occurred before additional surgery was performed 36 hours after the injury.

There were two instances of tamponade in these seven cases. One was entirely unsuspected until autopsy. The second was recognized too late for control, though it is doubtful that the patient could have survived, in view of the severe associated thoracoabdominal wounds.

The mode of death in the remaining (seventh) case in this group is not entirely clear. Death occurred suddenly and was considered due to myocardial dysfunction. The patient had a through-and-through wound of the left auricle, and there was approximately 100 cc. of blood in the pericardium, without any evidence of tamponade. It is speculative that repair of the defects might have been successful but quite obvious that a man with such injuries could not survive very long without surgery.


1. Makins, G. H.: Injuries to the Pericardium and Heart. In History of the Great War Based on Official Documents. Medical Services. Surgery of the War. London: His Majesty's Stationery Office, 1922, vol. I, pp. 431-475, passim.

2. Tribby, William W.: Examination of One Thousand American Casualties Killed in Action in Italy. Report to Surgeon, Fifth U.S. Army, 1944, 6 vols. [Official record.]

3. Elkin, D. C.: Wounds of the Heart. Ann. Surg. 120: 817-821, December 1944.

4. Bigger, I. A.: Heart Wounds. A Report of Seventeen Patients Operated Upon in the Medical College of Virginia Hospitals and a Discussion of the Treatment and Prognosis. J. Thoracic Surg. 8: 239-253, February 1939.

5. Wood, Paul.: War Wounds of the Heart. Proceedings of the Conference of Army Physicians, Central Mediterranean Forces, held at the Istituto Superiore Di Sanita Viale Regina Marguerita, Rome, 29 Jan. to 3 Feb. 1945, pp. 23-25.

6. Beck, C. S.: Further Observations on Stab Wounds of the Heart. Ann. Surg. 115: 698-704, April 1942.

7. Bright, E. F., and Beck, C. S.: Nonpenetrating Wounds of the Heart; A Clinical and Experimental Study. Am. Heart J. 10: 293-321, February 1935.

8. Warburg, E.: Myocardial and Pericardial Lesions Due to Non-Penetrating Injury. Brit. Heart J. 2: 271-280, October 1940.


9. Yates, John L.: Wounds of the Chest. In The Medical Department of the United States Army in the World War. Washington: Government Printing Office, 1927, vol. XI, pt. 1, pp. 342-442, passim.

10. Neurosurgery and Thoracic Surgery. Prepared and edited by the Subcommittees on Neurosurgery and Thoracic Surgery, Committee on Surgery, Division of Medical Sciences. National Research Council. Philadelphia and London: W. B. Saunders Co., 1943.

11. Cooper, F. W., Jr., Stead, E. A., Jr., and Warren, J. V.: The Beneficial Effect of Intravenous Infusions in Acute Pericardial Tamponade. Ann. Surg. 120: 822-825. December 1944.

12. Loison, cited by Lilienthal, Howard: Thoracic Surgery-The Surgical Treatment of Thoracic Disease. Philadelphia and London: W. B. Saunders Co., 1925, vol. I. p. 441.