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 VII



Management of Retained Foreign Bodies in the Heart and Great Vessels, European Theater of Operations

Dwight E. Harken, M.D.


Over a 10-month period in World War II, thoracic surgeons working in the thoracic surgery center set up in the 160th General Hospital, 15th Hospital Center, Cirencester, Gloucestershire, European Theater of Operations, U.S. Army, performed 134 operations for the removal of missiles retained in and about the heart and great vessels. There were no deaths in the series, and all patients were discharged with normally functioning hearts.

The experience of World War II confirmed the difficulty, commented upon elsewhere, of accurate localization of metallic foreign bodies in relation to wounds of the heart. In about half of the patients referred to the chest service at the 160th General Hospital as harboring foreign bodies within the heart, the objects were found, after careful fluoroscopic examination, to be extracardiac. Furthermore, early in the experience on this service, it was found at operation that about a third of the fragments thought to be intracardiac were extracardiac. Even on exploration, it was not always easy to determine whether an object lay within the pericardium or within the auricle; this was particularly true if an infected hematoma was present.

The location of the 134 foreign bodies in this series, as proved by operation, was as follows (fig. 144):

Of the 56 foreign bodies in and about the heart, 13 were intracardiac, 26 were entirely within the pericardium, and 17 partly within the pericardium but chiefly within the lung. Of the 13 foreign bodies removed from the chambers of the heart (fig. 145), 7 were in the right ventricle, 4 in the right auricle, l in the left ventricle, and 1 in the left auricle.

The great vessels were involved, directly or indirectly, in 78 cases. In seven instances, the missiles were intravascular; the three of these which became embolic are discussed in detail elsewhere (p. 380). In 35 instances, the fragments lay on the great vessels or within the vascular walls. In 17 cases, they were directly adjacent to the vessels but were chiefly within the lung. In 19


FIGURE 144.-Gross location of foreign bodies removed from cardiac and vascular structures at 160th General Hospital. Stippled: missiles impinging on structures. Black: intrastructural missiles. Crosshatching: embolic missiles.

instances, they were within the mediastinum but were not in direct proximity to the great vessels.

The fact that no missiles were encountered within the lumen of the thoracic aorta and that in only one instance was a missile found in the left ventricle seems to warrant the assumption that the direct entrance of foreign bodies of any considerable size into either of these areas was seldom compatible with survival. The anatomic distribution of these missiles suggests further that fragments small enough to enter the left ventricle as migratory foreign bodies were probably swept out of it by the high systemic pressure, in contrast to those which entered the right ventricle, where physiologic circumstances were more favorable for their lodgment.

Over the same 10-month period during which these 134 successful operations were performed, 4 unsuccessful operations were undertaken for the removal of foreign bodies in and near the heart. In two instances, the missiles


FIGURE 145.-Missiles removed from chambers of heart, with their location.

could not be recovered at cardiotomy, and in the other two instances, exploratory pericardiotomy revealed that their removal would have been attended with unjustifiable risks. 

In 15 other patients in whom foreign bodies in or about the heart were encountered, operation was not undertaken, for one reason or another, and the objects were permitted to remain in situ.


Indications for Surgery

The diagnostic confusion, already mentioned, of determining whether a retained missile was or was not within the heart added to the difficulties of assessing the risk to life and health of retained intracardiac missiles. In spite of numerous isolated reports of foreign bodies which have lodged in the heart


and remained asymptomatic, there is also considerable evidence in the medical literature that these objects may be lethal. Experimental evidence, furthermore, suggests that their presence may give rise to bacterial endocarditis.

As the war progressed, therefore, a working policy for the management of intracardiac missiles was formulated from the medical literature, from a limited amount of experimental work previously done on animals, and from the individual surgeons' own concepts and their accumulating experience. Although the policy was originally based upon somewhat uncertain premises, it was substantiated by the clinical evidence secured in the various chest centers in the European theater during the course of the war.

The indications for the removal of foreign bodies in and about the heart were chiefly based on possible risks, which included embolism, bacterial endocarditis, recurrent pericardial effusions, and myocardial damage. The development of symptoms and the existence of cardiac neurosis also served as indications.

The risk of embolism-The development of embolism from a foreign body or from the thrombus associated with it was not merely a theoretical possibility. Several such accidents had been recorded in the prewar literature, and at least two are known to have occurred in the European theater during World War II. One of these cases is included in the series analyzed in this chapter. The missile lay in the interauricular septum and right auricle and was removed from the auricle; the thrombus was found in the left auricle. Hemiplegia occurred in this case soon after wounding. In the second case, reported by Lt. Col. Arthur D. Nichol, MC, the embolism arose from a missile in the left ventricle. It is significant that the hemiplegia which followed it appeared more than 2 weeks after wounding.

The risk of infection-The removal of retained intracardiac objects because of risk of the development of bacterial endocarditis was, as already intimated, partly based on the results of experimental work on dogs. The implantation of foreign bodies in various locations in the heart was followed by the development of bacterial endocarditis characterized by typical bacterial valvulitis and of septic embolic infarcts. The experimental evidence was limited, but it was naturally feared that foreign bodies retained in the human heart might give rise to the same consequences.

Clinical support for this hypothesis was confined in this series to a single case: The patient ran a course suggestive of subacute bacterial endocarditis, with spiking fever, tachycardia, and one acute episode of pain in the right upper quadrant associated with jaundice. Response to surgical removal of the missile and the attached thrombus from the right auricle was immediate and dramatic. Although the patient was almost moribund when operation was undertaken, his recovery was prompt and uncomplicated.


The bacteriologic support for the indication of infection in this series of cases is somewhat more conclusive. No cultural studies were made in 4 of the 13 intracardiac foreign bodies. There was growth of bacteria on cultures made from only one of the five fragments removed from within the right ventricle; this particular foreign body lay in an abscess within a mural thrombus in the ventricular chamber. Cultures were also positive for pathogenic organisms in three of the four cases in which foreign bodies were removed from within the auricle. One of these missiles was found in a small abscess in the center of a mural thrombus in the right auricle.

It cannot be said positively that the infected niduses in these cases represented true bacterial endocarditis, nor can it be said that the infections present would eventually have produced it. Nonetheless, the findings seemed to furnish further support for the removal of intracardiac missiles because of the possible development of infection.

The risk of recurrent pericardial effusions-The risk of recurrent pericardial effusion as an indication for the removal of intracardiac foreign bodies has been repeatedly stressed in the medical literature and is well established. On the other hand, in the only two instances of pericardial effusion observed at the 160th General Hospital, the symptoms were not severe enough, and the fragments were not large enough, to justify surgical intervention.

The risk of myocardial damage-Myocardial damage was clearly evident in three cases in this series. The damage to the right ventricular wall overlying the site of a migratory missile noted in the following case seems of special significance in several respects. In particular, it demonstrates that the mere presence of a foreign body in a cardiac chamber can produce considerable damage to the overlying myocardium in a period of 3 months, and it also indicates that surgical removal of the object does not, of itself, cause myocardial injury of any consequence:

Case 1-A 29-year-old infantry sergeant was injured in the right lower posterior chest on 21 July 1944, in the fighting about Saint-Lô. Fluoroscopic and roentgenologic examination (fig. 146) showed a metallic foreign body lying in the anterior portion of the right ventricle, just to the left of the midline, and pulsating with the heart. An electrocardiogram 4 days later (fig. 147) showed no abnormalities except for inverted T-waves in the right-sided precordial leads (CF1, CF2, and CF3). By 8 August, the T-waves in CF3 had become upright, and the tracing appeared entirely normal.

First operation-Operation was performed on 15 August in the dorsal decubitus position (fig. 148). When the missile was grasped through an incision in the right ventricle, it was jerked from the grasp of the forceps by the movements of the myocardium and was lost to both sight and palpation in the bloodstream.

Roentgenograms taken immediately after the operation (fig. 149) showed the missile lying in the right auricle, over the opening of the inferior vena cava. Another electrocardiogram on 17 August (fig. 150) showed elevated ST segments in leads I and II, which fell by 1 September. Later, the T-waves also became sharply inverted in leads I and II and in the left-sided precordial leads CF4 to CF6. It was speculated that while this damage


FIGURE 146 (case 1).-Roentgenograms showing original position of fragment in right ventricle. A. Posteroanterior view. B. Lateral view.

FIGURE 147 (case 1).-Electrocardiograms taken before first cardiotomy. Roman numerals indicate limb leads and Arabic numerals precordial leads CF1 to CF6


FIGURE 148 (case 1).-Positioning of patient at successive cardiotomies. A. Dorsal decubitus (seen from above). B. Lateral decubitus (right side up). C. Reverse Trendelenburg position with elevation of right side.

might be related to the surgical incision made in the right ventricle near the septum, in the anterior surface of the heart, it might also result from an associated pericarditis.

Second operation-At a second cardiotomy on 16 November, 3 months after the first, with the patient in lateral decubitus (fig. 148), the missile was visualized and palpated in the right auricle, just above the entrance of the inferior cava. Again it escaped, this time falling back into the right ventricle (fig. 151). Incidentally, a significant diagnostic point is demonstrated by the roentgenograms taken after operation: They were not made in the true lateral position, and the impression was that the missile was in the chest wall. Had these been the first roentgenograms taken, considerable diagnostic confusion might have been caused.

Electrocardiographic examination (fig. 152) showed no specific acute change after the second cardiotomy. There was merely a progressive return toward normal of the T-waves and of leads I and II and CF5 and CF6.

Third operation-On 19 February 1945, at the patient's own request, a third cardiotomy was performed. The anterior approach was used, as in the first operation, but this time the patient was placed in the reverse Trendelenburg position (fig. 148). The scar

FIGURE 149 (case 1).-Roentgenograms showing position of fragment in right auricle after first (unsuccessful) operation. A. Posteroanterior view. B. Lateral view.


FIGURE 150 (case 1).-Electrocardiograms taken after first (unsuccessful) cardiotomy. See figure 147 for key.


FIGURE 151 (case 1).-Roentgenograms showing fragment again in right ventricle after second (unsuccessful) cardiotomy. A. Posteroanterior view. B. Lateral view.

of the first incision in the right ventricle, made 6 months earlier, was found solidly healed. This was demonstrated visually (plate I) as well as clinically by the colored motion pictures taken at the operation. The considerable fibrous pericarditis that had developed did not limit cardiac motion or obstruct the blood flow. Near the apex of the right ventricle, however, the muscle wall was thin, flabby, and discolored (plate II). The missile was palpable in the right ventricular cavity, just under the area of myocardial damage, which presumably had been produced by the friction of the muscle wall over the fragment during the 3 months after the second operation.

The heart was opened through the area of damage and the shell fragment (fig. 145) was grasped by forceps and was removed with only moderate difficulty (plate III). Intracardial manipulations were carried out in three episodes, over a period of about 3 minutes. Showers of extrasystoles were noted during the removal of the missile (fig. 153).

In spite of the successive migrations of this foreign body (fig. 154), this patient suffered no significant cardiac damage (fig. 155) and was in excellent condition (fig. 156) when he was discharged to the Zone of Interior.

FIGURE 152 (case 1).-Electrocardiogram taken after second (unsuccessful) cardiotomy. See figure 147 for key.


PLATE I (case 1) (right).-Appearance of heart at instant of incision into right ventricle (compare with figure 146A).

PLATE II (case 5) (left, bottom).-Technique of dislocation of heart from pericardial sac.

PLATE III (case 5) (right, bottom).-Marked ventricular dilatation following dislocation of heart. This photograph, like plate II, was taken during the operation. These plates should be compared with the electrocardiographic tracings taken during operation (fig. 173).


FIGURE 153 (case 1).-Electrocardiograms taken during third and successful cardiotomy at time of removal of fragment from right ventricle. See figure 147 for key.

FIGURE 154 (case 1).-Migration of foreign body at successive cardiotomies. A. Movement from right ventricle to right auricle as shown by dotted arrow; first operation in dorsal decubitus. B. Movement from right auricle back to right ventricle; second operation in left lateral decubitus, depicted by dotted arrow. C. Missile in right ventricle at third operation, in reverse Trendelenburg position, with elevation of right side. Solid arrow designates site of successful surgical removal of foreign body through anterior wall of right ventricle.


FIGURE 155 (case 1).-Electrocardiograms taken after third and successful cardiotomy with removal of fragment from right ventricle. See figure 147 for key.

FIGURE 156 (case 1).-Patient just before evacuation to Zone of Interior, after three cardiotomies for removal of retained intracardiac missile.


FIGURE 157 (case 2).-Electrocardiograms taken before cardiotomy showing left ventricular damage from retained missile in left ventricle. See figure  147 for key.

Case 2-Another instance of myocardial damage occurred in the only case in the series in which the foreign body was in the left ventricle. Roentgenokymographic studies showed diminished amplitude of pulsation at the apex and passive left ventricular dilatation during systole, suggestive of an early ventricular aneurysm or hernia. Electrocardiographic studies (fig. 157) showed a persistent pattern characteristic of extensive damage to the anterior left ventricular wall, consisting of low voltage, deep Q1, absent R1, and inverted T1, together with inverted and W-shaped QRS complexes and sharply inverted and coved T-waves in the left-sided precordial leads.

At operation, the foreign body was found in the left ventricle, as the roentgenograms (fig. 158) had shown, in a cystic zone of myocardial damage 1.5 cm. in diameter. It was ballotable in the defect in the cardiac wall, and parodoxical pulsation of this area of the ventricle was noted. A mural thrombus, which was not disturbed, permitted removal of the missile without hemorrhage. The myocardial defect, after closure, was covered over with two superimposed pericardial grafts.

In this case, electrocardiographic tracings taken at frequent intervals during the operation showed no evidence of cardiac irritability at any time except for a few ventricular extrasystoles during the process of endotracheal intubation. It was believed that removal of the missile and repair of the myocardial defect prevented progression of the myocardial damage and averted possible rupture of the heart. Direct inspection at operation left no


FIGURE 158 (case 2).-Preoperative roentgenograms showing retained missile in left ventricle. A. Posteroanterior roentgenogram. B. Lateral roentgenogram.

doubt that the risk of rupture of the myocardial hernia was real and that it was aggravated by the presence of the foreign body in it.

Case 3-A third instance of damage to cardiac structures by a retained foreign body was observed by Maj. Fred J. Jarvis, MC. In this instance, the wall of the right ventricle underwent such a degree of degeneration from the presence of the underlying migratory missile that death ensued.

Other indications for surgery-Two additional factors sometimes assumed importance in the decision to remove intracardiac foreign bodies. One was pain, the other cardiac neurosis.

Pain was associated with some of the pericardial foreign bodies in this series, but with only one of the intracardiac missiles. In this case, the object had migrated from the auricle to the ventricle. A similar case was observed by Lt. Col. (later Col.) Laurence Miscall, MC.

Cardiac neurosis sometimes became a pressing indication for surgery. In spite of every effort to reassure them, all the patients in this series with foreign bodies in or near their hearts wanted them removed. Their apprehensions bore out Grey Turner's (1) remark, "In addition to the characteristic cardiac symptoms * * * there may be neurotic manifestations which mainly depend on the attitude of the patient to the knowledge that he harbors a foreign body in one of the citadels of his well-being."

Size and location were other factors which influenced the decision for and against surgical intervention. Small foreign bodies left in situ were regarded as less hazardous than larger ones, and it was also thought that they were associated with less risk of myocardial damage. An additional reason for leaving them undisturbed was that they became encapsulated more readily and more firmly than larger missiles and were therefore technically more difficult


FIGURE 159 (case 4).-Preoperative roentgenograms showing retained fragment in left auricle. Radiopaque oil is seen in empyema pocket. A. Posteroanterior view. B. Lateral view.

to remove. For these reasons, small, silent foreign bodies in and near the heart were allowed to remain in situ in 15 cases at the 160th General Hospital during the period in which 13 similar fragments were removed surgically.

Combined indications-As might have been expected, more than one indication for surgery was sometimes present, as the following case history demonstrates:

Case 4-This soldier, following injury by a shell fragment in the left anterior aspect of the chest, developed an empyema, from which both hemolytic Staphylococcus aureus and Clostridium welchii were cultured. He had been treated by decortication followed by open drainage at the 160th General Hospital. In the 6 months after injury, he had three massive episodes of bleeding and two minor episodes. He also had bouts of pyrexia to 103° F., which did not appear to be connected with his empyema. Posteroanterior and lateral roentgenograms (fig. 159) revealed a shell fragment in the left auricle. Electrocardiographic studies before operation (fig. 160) showed right axis deviation with low R1 and deep S1, low diphasic T1, and upright, pointed T2 and T3. The precordial leads were normal. The abnormalities suggested damage to the anterior wall of the myocardium but were not helpful in localizing the missile.

At operation (fig. 161) on 18 May 1945, the empyema was found to communicate with a laceration in the pericardium and the underlying left auricle; the auricle was adherent to the pericardium. A laceration in the auricle was plugged by a large, infected, intracardiac hematoma. The missile, which was 2 by 1 by 1 cm. and was surrounded by cloth, lay in this clot. It was readily removed from the left auricle, together with the cloth and the infected clot. Cl. welchii and Escherichia coli were grown on direct culture from the material removed.

After operation, the empyema rapidly resolved and there were no further episodes of hemorrhage or pyrexia. Serial electrocardiograms (fig. 162) showed deep and sharp inversion of T1 and inverted T in CF6 but no other significant change.

This case appears to embrace most of the indications for removal of an intracardiac foreign body (including an extensive thrombus), a potential embolus, gross intracardiac contamination and infection, pericardial involvement, and damage to the myocardium with repeated hemorrhages.


FIGURE 160 (case 4).-Electrocardiogram taken before operation showing possible damage to anterior myocardial wall. See figure 147 for key.

FIGURE 161 (case 4).-Technique of removal of retained shell fragment from left auricle. A. Patient in right lateral decubitus. B. Laceration of left auricle plugged by infected hematoma. C. Closure of laceration by pericardioauricular sutures.


FIGURE 162 (case 4).-Electrocardiogram after left auricular cardiotomy and removal of intra-auricular shell fragment. See figure 147 for key.

Principles of Surgery

A successful attempt to remove a foreign body from the heart depended upon the observance of certain fundamental principles governing exposure of the involved area and manipulation of the heart itself. The most important of these principles were:

1. Adequate, direct exposure of the affected area. This necessitated the use of a variety of approaches (fig. 163), depending upon the special problem at hand. The pleura was opened routinely, but no standard or inflexible cardiac approach was relied upon.

2. Conservation of the skeleton of the thoracic cage. The free division of ribs, cartilage, and sternum was often necessary to reach the heart. The division of bone and cartilage did no harm, but it was essential that these structures be preserved in situ. Complete reconstruction of an intact chest wall was necessary at the end of the operation. No tissue or bone was resected, and nothing was discarded. As a result, there was neither deformity nor defect at the end of the operation.


FIGURE 163.-Incisions used to obtain exposure of foreign bodies in heart. A. Anterior aspect, showing: Approach to superior vena cava and right auricle (a), and approach to anterior and lateral aspects of right and left ventricle (b). B. Posterior aspect, showing: Approach to posterior aspect of right left ventricle (a), and approach to posterior aspect of right and left auricles (b).

3. Maintenance of a moist epicardium during the period of cardiac exposure. It was thought that the 1 percent procaine hydrochloride solution used to keep the epicardial tissues moist might also have the additional advantage of reducing cardiac irritability.

4. Minimal dislocation of the heart from the position of optimal function. To perform cardiac surgery with minimal cardiac dislocation often taxed the ingenuity of the surgeon, but it was essential for a successful result.

Surgical Technique

Both the approach and the procedure for the removal of foreign bodies in and about the heart depended upon their location. The various approaches described were all used successfully in one or more cases in this series. Another extracardiac approach to the heart, by way of the pulmonary veins to the left auricle or even the left ventricle, was tentatively discussed, but no occasion to use it arose.

Right ventricle-When the object was in the right ventricle (fig. 164), the best approach was through the fifth or the fourth intercostal space (fig. 165), which exposed the left portion of the right ventricle and the left border of the left ventricle. The pectoralis major was split laterally and divided medially. The intercostal muscles were divided, and the internal mammary vessels were ligated. The cartilage superior to the incision was divided near the sternum, after which the rib-spreader was inserted. If additional exposure


FIGURE 164.-Roentgenograms showing foreign body in chamber of right ventricle. Before these films were made, the right hemithorax had been cleared of an infected hemothorax. A. Posteroanterior view. B. Lateral view.

FIGURE 165.-Technique of removal of foreign body shown in figure 164. A. Skin incision over the anterior left fifth interspace with optional extension. B. Bony cage exposed: fifth and sixth cartilages divided (a), fifth intercostal space incised (b). C. Exposure of right ventricle by incising fused pericardium (a), and exposing foreign body (b). D. Placement of hemostatic sutures in right ventricular wall and removal of foreign body. E. Enlargement demonstrates crossing of hemostatic sutures.


FIGURE 166.-Technique of combined intercostal-transsternal approach to right aspect of right ventricle and phrenic surface of heart and pericardium. A. Skin incision. B. Transverse section of sternum with Gigli's saw (a), incised left fifth intercostal muscle (b). C. Exposure of foreign body with aid of Tudor Edwards' double retractor: foreign body (a), localized abscess of pericardium (b). D. Digital stabilization of dislocated heart.

was necessary, it could be obtained by the optional T-incision and division of the sixth and seventh costal cartilages.

After the pericardium and pleura had been opened widely, this approach permitted palpation of a foreign body in the right ventricle. As a rule, the cardiac chamber was not opened, nor was any attempt made to remove the missile, until it had been located by palpation. When it had been located, two rows of sutures were placed immediately over the object, on either side of the projected incision. These sutures, which served as hemostatic guy or control sutures, were used by the assistant to prevent blood loss from the opened heart between intracardiac maneuvers. The second row of sutures also served as a


FIGURE 167.-Roentgenograms showing missile in right auricle. A. Lateral view. B. Posteroanterior view.

sort of second line of defense, in the event that any of the sutures in the first row pulled out.

After the missile had been stabilized as well as possible with the fingers, a small epicardial incision was made with the scalpel, and a pointed forceps was thrust through the myocardial wall and spread open; an opening sufficiently large to permit the insertion of a Kocher clamp was thus created. The foreign body was then grasped and extracted with this instrument. During these manipulations, the assistant crossed the control sutures and thus produced hemostasis.

Once the fragment had been removed, the inner row of control sutures was tied across the incision. The second line of sutures was then tied over a small free pericardial graft, over which a second, larger, free pericardial graft was sutured into place. Postoperative cardiac tamponade was avoided by leaving the lateral angle of the pericardial incision open into the pleural space.

Chromic catgut sutures No. 00 were used, on atraumatic noncutting needles. In retrospect, it might have been wiser to use silk for the ventricular closure.

Diaphragmatic surface of the heart-When the foreign body lay farther to the right or was on the diaphragmatic surface of the heart, the intercostal incision described was combined with a transverse sternal section (fig. 166). Tudor Edwards' double retractor could be usefully adapted for this procedure. The sternum was readily reconstituted with wire.

When the auricle or the right side of the ventricle had to be approached, the reverse of this incision could be used by following precisely the same technique on the right side of the chest.

Right auricle-When the missile was in the right auricle (fig. 167), a right anterior approach was usually best (fig. 168). The third, the fourth, or even the fifth interspace was used, depending upon the location of the foreign body. The third interspace, for instance, was incised, with section of the


FIGURE 168.-Technique of removal of foreign body shown in figure 167. A. Incision for anterior approach to right atrium. B. Closure of incision with pericardioatrial and atrial sutures. C. Exposure of right atrium (a), reflected pericardium (b).

third and fourth costal cartilages near the sternum, to provide exposure of the upper portion of the right auricle. After the pericardium had been opened, control sutures were placed in the auricular wall by the technique described for their application in the ventricular wall. The auricular myocardium was often too thin to hold sutures. If the structures were adherent, as in the case illustrated, pericardial tissue could be included with auricular tissue in the second suture line.

Left auricle-If the foreign body was in the left auricle, the approach shown in figure 161 was used. In this case (case 4, p. 367), the laceration of the auricle and pericardium was plugged by an infected clot, and the fusion of the two structures made the intracardiac portion of the operation perfectly simple.

Pericardium-In this series, intrapericardial missiles were removed by a posterior approach that exposed the left auricle on either the right or the left side. This technique was not used in any intracardiac operation at the 160th General Hospital. When the patient was placed in ventral decubitus (fig. 169), the posterior incision provided adequate exposure.


FIGURE 169.-Technique of posterior approach to pericardium and heart on right. A. Incision with patient in ventral decubitus. B. Exposure.

In one case in which this approach was used (fig. 170), an extremely thin-walled auricle was encountered in a completely free pericardial space. Bleeding from the auricular wall was so free when an attempt was made to place the hemostatic sutures that this technique had to be abandoned. An approach through the auricular appendage proved equally unsafe. An extracardiac approach was therefore devised, by exposure of the superior vena cava (fig. 171). Control tapes were placed around this vessel, an incision was made into it, and a forceps was slipped down the lumen and into the ventricle. This maneuver permitted removal of the missile without difficulties.

Behavior of the Heart During Cardiac Surgery

The electrocardiographic tracings made at operation in some cases in this series were useful in demonstrating manipulations that were not well tolerated by the heart. These manipulations included cardiac dislocations, hemostatic cardiac grips, and extensive maneuvers within the cardiac chambers.


FIGURE 170.-Roentgenograms showing foreign body in interauricular septum and right auricle. The widened mediastinum was found at operation to be caused by a hematoma. A. Posteroanterior view. B. Lateral view.

Cardiac dislocation-In some instances, dislocation of the heart at operation caused runs of ventricular extrasystoles, diminished cardiac output, and transitory bundle branch block. Whether the dislocation was accomplished manually or by means of an apical suture, it was not well tolerated, as the following case history shows:

Case 5-At operation, this foreign body (fig. 172) was found extracardially, in a pericardial abscess well back on the diaphragmatic surface of the heart. The abscess contained about 18 cc. of pus. To gain access to this area, the heart had to be lifted out of the pericardial sac (plate II). This maneuver caused a fall in blood pressure, accompanied by circulatory failure, which made it necessary to replace the heart frequently, after relatively short periods of dislocation, to permit a return to normal conditions before the operation could be proceeded with. Many irregularities in rhythm occurred, which were apparently extrasystoles, and marked cardiac dilatation, particularly of the right ventricle, also developed, so that the heart became too large for the pericardial sac (plate III). After it was replaced, the pericardium was left partly open.

Electrocardiograph tracings taken during operation (fig. 173) showed variations in rhythm consisting of ventricular extrasystoles (at 300), wandering pacemaker, varying P-R interval, and A-V nodal rhythm. It was at the time the nodal rhythm occurred, during a period of dislocation of the heart, that a particularly prolonged and alarming episode of circulatory failure developed. After recovery, normal sinoauricular tachycardia returned (fig. 173).

An additional change in the electrocardiograms was also related to dislocation of the heart. At 2:57 o'clock (fig. 173), the S-wave became broad and notched and the QRS interval lengthened to 0.13 second, in contrast to the normal complexes at 2:07 and 2:37 before cardiac manipulations. While it is unfortunate that this abnormality was recorded only in lead II, it may be regarded as indicating at least interventricular block or incomplete bundle branch block, probably on the right side. The abnormal QRS complexes persisted throughout the operation, but 4 days later, observations were normal. At operation,


FIGURE 171.-Technique of removal of foreign body in right auricle by incision in superior vena cava shown in figure 170. A. Anterior skin incision. B. Third intercostal space incision (a), section of third and fourth costal cartilages close to sternum (b), and exposure of intercostal vessels to permit their division (c). Excellent exposure of superior vena cava by this approach: pericardium incised (a). D. Incision into vessel and removal of right auricular foreign body (a), tapes placed around superior vena cava to control hemorrhage (b). This technique was used in this case because the pericardial space was free and the auricular wall too thin to hold guy sutures.

the delay in conduction could be correlated, by direct visual inspection, with dilatation of the right ventricle. This extremely unusual observation in a human subject was considered to be caused, at least in part, by the increased time necessary for conduction of the impulse through the greatly dilated right ventricle.

The intolerance of the heart to dislocation was demonstrated in this case in two ways, (1) by ventricular dilatation with incomplete bundle branch block, and (2) by varying types of arrhythmia and circulatory collapse. Obviously, dislocation of the heart may produce torsion of the great vessels and obstruction to outflow of blood, with (1) a fall in blood pressure resulting from the diminished cardiac output and (2) ventricular dilatation resulting from the increased resistance to blood flow.


FIGURE 172 (case 5).-Roentgenograms showing foreign body in pericardial abscess well back on diaphragmatic surface of heart. A. Posteroanterior view. B. Lateral view.

This sort of experience led surgeons on the thoracic surgery service at the 160th General Hospital to avoid the apical suture as a means of exposing inaccessible areas of the heart. The situation encountered in combat-incurred wounds is quite different from the circumstances of civilian surgery, in which a planned approach permits precise and comfortable exposure of any part of the heart.

Obstruction of the cardiac blood flow-Experience showed that the classical hemostatic cardiac grips intended to provide a bloodless surgical field were also poorly tolerated. These maneuvers disturbed cardiovascular dynamics and were therefore used only as a last resort.

The intolerance of the heart to obstruction of the blood flow was in striking contrast to its stability during other cardiac procedures and certain intracardiac procedures. The surface of the heart could be manipulated, sutures taken into the muscle, and incisions made into the cardiac chambers with little evidence of disturbance.

Intracardiac manipulations-More extensive manipulations inside the cardiac chambers by the exploring finger or by a forceps inserted to remove an intracardiac thrombus or a foreign body were less well tolerated. Sometimes no abnormalities were noted, but marked cardiac irregularity, in the form of multiple ventricular extrasystoles, were common. The case, already described (case 1, p. 357), in which three cardiotomies were necessary, demonstrates this point particularly well. As the foreign body was grasped and extracted from the right ventricle, electrocardiographic tracing (fig. 153) showed showers of ventricular extrasystoles from different foci in both ventricles; runs of ventricular tachycardia lasted as long as 16 seconds. Direct observation of the irregular heart action and examination of the electrocardiogram raised the


FIGURE 173 (case 5).-Electrocardiograms taken during operation for removal of pericardial foreign body. See figure 147 for key. The number in the lower corner of each segment indicates the clock reading at which the tracing was made. Note the nodal rhythm, which occurred during a prolonged and alarming episode of circulatory failure while the heart was dislocated (plate II).

fear of impending ventricular fibrillation, but at the end of the operation, when the irritating forceps and the missile had been removed from the ventricular chamber, the tachycardia ended promptly and the P-R interval returned to normal in three beats. All of these irregularities were relatively benign.

Other causes of cardiac irritability-Minor evidences of cardiac irritability were observed at operation, such as extrasystoles and wandering pacemaker, varying P-R interval, and even A-V nodal rhythm, but these phenomena were not usually accompanied by any significant clinical manifestations. These minor abnormalities were often evoked by procedures not related to the heart, such as endotracheal intubation, spreading of the ribs, and manipulations of the hilar and mediastinal structures. It was thought that keeping the surface of the heart moist with warm physiologic salt solution or procaine hydrochloride solution was important in reducing the tendency toward irritability.


Postoperative Management

The postoperative care of a patient who had undergone cardiac surgery was essentially the same as the care of a patient who had undergone any serious chest surgery. Patients who had required ventricular cardiotomy were kept in bed for 3 weeks, even though the electrocardiographic tracings had usually returned to normal within a few hours after operation.


Indications for Surgery

The danger of erosion and suppuration as the result of the retention of large missiles in close proximity to thoracic blood vessels was quite real. At least three deaths are known to have occurred at other installations in the European theater as the result of erosion from this cause. In the 78 successful operations for removal of such missiles at the 160th General Hospital thoracic surgery center, there was no instance of erosion. On the other hand, abscess formation occurred in about 15 percent of the cases; clothing, bone fragments, and other foreign material entered the body with the shell fragment in 30 percent of the cases; and pathogenic bacteria were found on 67 percent of the material that was cultured on removal.

On the basis of these facts, it became established policy at this chest center to remove all foreign bodies in the mediastinum adjacent to the great vessels if they measured 1 cm. or more in two dimensions. The decision to remove smaller objects or permit them to remain in situ was made on the indications in each case.

Removal of the foreign bodies seldom presented any technical difficulties. At the 160th General Hospital, operation was unsuccessful in only one case, the third undertaken. The fact that there were no deaths in the 78 cases in which the objects were removed seems to support the contention that their removal was safer than their continued retention.

The only cases in this series which warrant special discussion are the three instances of migratory intravascular objects, which will be described in some detail.


In 1942, Straus (2) collected from the world medical literature 32 instances of migratory intravascular foreign bodies and added to the collection a case which he had himself observed. A second review of the same literature in 1945 brought to light nine additional cases that Straus did not include in his material. Undoubtedly, other cases were overlooked in both reviews, while still other cases have not been reported at all.


FIGURE 174 (case 6).-Roentgenograms showing .30-caliber bullet in hilar region of left lung. The bullet was later found to be within one of the radicles of the left pulmonary artery. A. Posteroanterior view. B. Lateral view.

Almost without exception, the migratory foreign bodies in these 42 cases were bullets or shell fragments, the majority of which originated in combat-incurred injuries.

The infrequency of such cases, the bizarre courses of many of the embolic missiles, and the dramatic developments often associated with their migration warrant a presentation of the three such cases encountered at the 160th General Hospital thoracic surgery center, together with a brief comment on certain of their features.

Case 6-A 30-year-old soldier was wounded in the left infraclavicular region by a .30-caliber machinegun bullet while in combat in Belgium on 26 December 1944. The wound was debrided and sutured the same day at a forward installation. Roentgenograms of the chest taken at this time revealed that the bullet was located posterior to the hilus of the left lung.

When the patient was evacuated to the 128th General Hospital, Bishopstrow, England, it was decided that the size of the missile and its proximity to the hilus of the lung constituted valid indications for its surgical removal. He was therefore transferred to the chest center at the 160th General Hospital on 8 January 1945.

Initial roentgenograms (fig. 174) confirmed the previous findings. Routine fluoroscopic studies for precise localization of the missile suggested nothing more remarkable than the presence of a pulsating bullet intimately associated with the left hilar structures. There seemed no doubt that it could readily be removed through a conventional posterolateral approach.

Thoracotomy was accordingly performed through this approach on 19 January 1945. After resection of 12 cm. of the seventh rib, the bullet could be palpated in the central portion of the left upper lobe. Its location was confirmed by palpation by the first assistant. When, however, the surgeon palpated the area for a second time, he was chagrined to find that the missile had vanished. A thorough search of the entire left lung, including the hilus, and of the posterior mediastinum and the pericardium failed to locate the bullet, and it was suspected that it had fallen into the tracheobronchial tree. When this possibility


FIGURE 175 (case 6).-Roentgenograms taken immediately after left thoracotomy. The bullet now seems to be in the hilus of the left lung. The double image of the bullet in the lateral view is explained by its pulsation with the pulsation of the pulmonary artery. A. Posteroanterior view. B. Lateral view.

was excluded by immediate bronchoscopy, no further manipulations were attempted, and the chest was closed.

Recovery was completely uneventful and the patient was allowed out of bed daily after the first 24 hours. Roentgenograms taken immediately after operation (fig. 175) showed that the bullet that had previously been seen in the hilar region of the left lung (fig. 174) was now lying horizontally, just to the right of the body of the sixth dorsal vertebra, pointing outward. The pulsations indicated that it was in contact with the area of the pulmonary artery. It was concluded that the missile was definitely in the right pulmonary artery.

A second thoracotomy was performed on 20 February 1945, through the third intercostal space anteriorly (fig. 176). The third and fourth costal cartilages were divided close to the sternum. There was no evidence of altered blood flow in the lungs. The bullet was palpated in the hilus of the right lung at and in the mediastinum, posterior to the superior vena cava.

The right pulmonary artery was isolated, and a tape was placed about it proximal to the foreign body. The missile, point outward, was digitally manipulated away from the hilus into the pulmonary artery of the upper lobe, the milking process being continued until its point presented beneath the visceral pleura in the interlobar fissure. Here, the pleura was nicked, and the presenting tip was thrust upward through the opening and grasped with a Kocher clamp. The missile was then delivered in toto by slowly stretching the terminal branches of the pulmonary arterial tree. Hemorrhage was controlled by circumferential mattress sutures of catgut in the lung at the point of removal of the missile. The wound was closed without drainage.

The patient was up and about after the first postoperative day. His recovery was completely uneventful, and he was returned to the Zone of Interior for rehabilitation on 30 March 1945.

Case 7-This soldier was wounded on 14 October 1944, in action in France, by a shell fragment that entered through the upper part of the right posterior axillary line. The wound of entrance was debrided and closed the same day at the 60th Field Hospital. Aspiration of a right hemothorax yielded 500 cc. of blood. Roentgenograms of the chest (fig. 177) revealed a foreign body in the right upper mediastinum, 3 cm. posterior to the right sternoclavicular joint.


FIGURE 176 (case 6).-Technique of removal from right pulmonary artery of foreign body shown in figures 174 and 175. A. Skin incision. B. Exposure of right pulmonary artery. Umbilical tape placed proximally around the main pulmonary prevents escape of the bullet to opposite pulmonary artery or right ventricle. C. Milking of bullet to periphery of lung, into one of the smaller radicles of the pulmonary artery. D. Opposing hemostatic sutures at the site of removal of bullet.

After the patient reached the 154th General Hospital, Wroughton, England, a pericardial friction rub was heard. On 21 October, a bruit was discovered over the right supraclavicular fossa, and it was noted that both the brachial and the radial pulses were absent in the right arm. Thoracentesis, repeated on 7 November, yielded what was at first thought to be fresh blood. It was later found to be old, changed blood of the so-called grapejuice type.

In view of the roentgenologic and clinical findings, the patient was transferred to the thoracic surgery center at the 160th General Hospital on 10 November 1944. There, it was noted that the right hand was cooler than the left and was waxy-yellow. The blood pressure in the right arm was 102/80 and in the left 122/68 mm. Hg. Fluoroscopic and roentgenologic examination led to the conclusion that the foreign body lay in close proximity to the innominate artery.

Surgical removal of the fragment was considered advisable because of (1) its size; (2) its proximity to the large vessels in this region; (3) the evident vascular damage it had caused, as attested by the bruit and the differences between the blood pressures in the right and left arms; and (4) the possibility that fresh bleeding had occurred into the right pleural cavity. The latter consideration suggested that surgical intervention be instituted promptly.


FIGURE 177 (case 7).-Roentgenograms showing retained shell fragment posterior to right sternoclavicular joint in region of innominate artery. A. Posteroanterior view. B. Lateral view.

Operation was performed on 12 November 1944 under endotracheal gas-oxygen-ether anesthesia. A U-shaped incision was made over the right upper anterior chest (fig. 178). The sternum was exposed. The internal mammary vessels were divided and ligated in the second intercostal space. The right pleural cavity was opened widely. The deep surface of the sternum was cleared of underlying structures, and the bone was split with a power saw longitudinally in the midline, from the suprasternal notch down to the level of, and out through, the second intercostal space. When the segment of sternum, with the clavicle and the first two ribs attached, was reflected upward and outward, excellent exposure of the superior mediastinum and the base of the neck was secured.

The superior vena cava and the innominate veins were retracted with tapes. The innominate artery was isolated at its point of origin from the aorta, and a tape was placed about it for hemostatic purposes. It was then dissected free up to its bifurcation into the carotid and subclavian arteries. Tapes were placed about each of these vessels. It was not possible to palpate the foreign body within the upper portion of the innominate artery, the wall of which was intact and free from scars. The artery was incised longitudinally, and the foreign body was removed. The portion of intima which had been in contact with it showed evidences of patchy destruction. After thrombi had been removed from the lumen of the vessel, bleeding was free, both proximally and distally.

The incision in the artery was closed in two layers, with interrupted No. 000 silk sutures. The first layer included the entire wall. The second sutures were introduced through the adventitia and media; a bite was taken on either side of the incision as each suture was placed, but the sutures did not pass through the lips. Fibrin foam was laid over the incision and held in place with the second layer of sutures.

The sternum was approximated with three interrupted wire sutures placed in drilled holes that had been so staggered that the medial holes were lower than the lateral holes. When the sutures were pulled tight, the sternal fragment was therefore drawn downward and inward, and a precise and stable reconstruction of the pectoral girdle was thus obtained. Great care was taken to accomplish this result because stability had been found to be of great importance after section of the sternum; the movement of loosely approximated sternal fragments was painful and could produce shock.

The pectoral muscles were approximated with interrupted sutures of No. 00 chromic catgut. The superficial fascia was similarly closed, and the skin was sutured with No. 0 black silk. No drains were used.


FIGURE 178 (case 7).-Technique of removal of retained missile from innominate artery. A. Incision. B. Sternal-splitting approach. C. Detail of position of foreign body in innominate artery (a), ligated thyroid vein (b), and right vagus nerve (c). D. Detail of hemostasis before removal of missile and repair of artery. E. Detail of technique of vascular closure is shown in (a) and (b), fibrin foam used to reinforce vascular closure represented by (c).


FIGURE 179 (case 8).-Roentgenograms 3 days after wounding showing clear lung fields. A. Posteroanterior view. B. Lateral view.

Immediately after operation, the pulse in the right arm was barely perceptible. The first postoperative day, the blood pressure was 122/80 mm. Hg in the left arm but remained unobtainable in the right arm. The second postoperative day, the right stellate ganglion was infiltrated with procaine hydrochloride. The right arm immediately became warm, and a systolic blood pressure of 68 mm. Hg could be obtained. On 17 November, the blood pressure was 132/60 mm. Hg in the left arm and 98/76 in the right arm. On November 25, the respective pressures were 124/86 and 108/64 mm. Hg. On 15 January 1945, the blood pressures were the same in both arms.

The patient was heparinized for 3 days after operation. The clotting time usually averaged from 9 to 15 minutes but on a single occasion was 45 minutes. The hemothorax still present after operation was evacuated by aspiration and required no further treatment. Hoarseness, the result of manipulation of the right recurrent laryngeal nerve, was present for 4 weeks after operation; then it disappeared completely. Union of the sternum was firm. When the patient was returned to the Zone of Interior on 1 March 1945 for rehabilitation, he was in excellent condition and had no evidence of any circulatory disturbance.

Case 8-A 22-year-old soldier was struck by a machinegun bullet on 15 July 1944, while in action in France. The point of entrance was in the right axilla, at the level of the fifth rib. When he was evacuated to the 159th General Hospital, Yeovil, England, on 17 July, routine roentgenograms showed no abnormalities (fig. 179), and the lateral film established the absence of metallic objects in the retrocardiac zone.

The patient was symptom-free when he was sent to the rehabilitation barracks, and for about 3 weeks he participated in a vigorous athletic program, with no difficulties. Then, early in the morning of 13 August, he was wakened by agonizing pain in the lower left chest, accompanied by dyspnea, orthopnea, cough, and hemoptysis. Rales and diminished breath sounds were heard over the left axilla. On 15 August, roentgenograms of the chest (fig. 180) revealed a .30-caliber bullet in the hilus of the left lung. On 19 August, bronchoscopy revealed edema of the left main bronchus with an injected bleeding point on the posterolateral wall. The continued hemoptysis and the bronchoscopic findings suggested that the bullet had eroded from the pulmonary artery into the left main or descending bronchus. The patient was therefore transferred to the chest unit at the 160th General Hospital on 22 August.


FIGURE 180 (case 8).-Roentgenograms 1 month later, 2 days after sudden outset of pain in left lower hemithorax. A .30-caliber bullet is now visualized in the hilar region of the left lung. At operation, it was found in the left inferior branch of the pulmonary artery. A. Posteroanterior view. B. Lateral view.

Since the possibility of erosion of the bullet from the pulmonary artery into the bronchus seemed an urgent indication for surgical intervention, operation was undertaken on 28 August 1944. Thoracotomy was performed through a posterolateral approach, with resection of a 16-cm. segment of the left seventh rib. Exploration revealed discoloration of the lingular portion of the left upper lobe and, to a lesser degree, of the anterior basic segment of the left lower lobe. The missile could be palpated through the interlobar fissure in the lower lobe and hilus; it seemed intimately associated with the artery. A fluctuant area, about 1 cm. in diameter, was noted over the base of the bullet.

The hilus was exposed, so that the pulmonary artery could be compressed by grasping the hilar structures from above downward between the index and middle fingers of the left hand. The digital compression not only was hemostatic but also served to stabilize the missile. With the situation thus under control, the fluctuant area was incised. The butt of the bullet was found lying in an abscess cavity anterior to, and including the anterior surface of, the lower branch of the left pulmonary artery. The body and point of the missile, however, lay within the arterial lumen, and as soon as the bullet had been extracted, the full arterial stream issued from the defect. The hemorrhage was at all times completely under control, as could be proved by voluntary release of the artery and immediate cessation of the bleeding when compression was resumed.

The abscess cavity was thoroughly debrided, and mattress sutures were placed through its wall and through the subjacent artery. Fibrin foam impregnated with penicillin was placed over the mattress sutures used to close the arterial incision before they were tied. An excellent closure of the vessel was thus effected. The gangrenous lingular portion of the left upper lobe was then resected, and the line of resection was closed with mattress sutures of No. 00 chromic catgut. The anterior basic segment of the left lower lobe, although somewhat discolored, was considered viable and was not disturbed. The lung reexpanded readily. Before the chest was closed, 100,000 units of penicillin in 50 cc. of 2.4-percent sodium citrate solution were instilled into the pleural cavity.

The left hemothorax present after operation was aspirated and did not recur. The patient was up and about the ward within 48 hours, and he was in excellent condition when he was transferred to a rehabilitation center on 1 December 1944.


FIGURE 181 (case 6).-Migration of bullet from left pulmonary artery into main pulmonary artery and thence into right pulmonary artery: Original position of bullet in left pulmonary artery (a), and final position of bullet after migration into right pulmonary artery (b).

Migration and Clinical Consequences

Pattern of migration-In the 42 cases of migratory intravascular foreign bodies reported in the literature up to 1945 (p. 380), migration was about equally divided between arterial and venous channels, depending upon which the object happened to enter at the time of wounding.

Apparently, there is considerable latitude in the possible patterns of migration when a foreign object enters the circulation.

In 11 of the 42 recorded cases, the missile entered one of the larger arteries and passed peripherally as an embolus. In five instances, it entered the left ventricle and was swept out into the aorta, to lodge within this vessel or within one of its branches. In one case, it passed through the left ventricle, to become embolic to the right femoral artery. Two missiles entered the ascending aorta and passed downward against the arterial stream; one came to rest in the sinus of Valsalva and the other in the left ventricle. One missile entered the pulmonary artery, migrated downward into the right ventricle, and killed the patient 7 days later, when it became embolic to the right pulmonary artery (3).


FIGURE 182 (case 7).-Probable course of migratory missile from left auricle through left ventricle and aorta into innominate artery: Foreign body enters right pulmonary vein (a), left atrium (b), left ventricle (c), and innominate artery (d).

In two instances, the missile entered one of the pulmonary veins and passed into the left ventricle.

The patterns of arterial migration described in two of the three cases of migratory foreign bodies observed at the chest center at the 160th General Hospital do not seem to have been reported previously. In one case (case 6, fig. 181) migration was from the left pulmonary artery into the main pulmonary artery and thence to the right pulmonary artery. The bullet was evidently in the lumen of the left pulmonary artery at the first operation and was dislodged by the exploratory palpation.

In one case (case 7, fig. 182), since the wall of the innominate artery and the adjacent portion of the aorta were found at operation to be intact everywhere, the course of the missile, it seems, must necessarily have been through the right side of the chest and into the pulmonary vein or the left side of the heart. The fragment was then carried as an embolus by the bloodstream into the aorta and thence to the innominate artery, where it lodged. The pericardial friction rub noted before operation suggested that the pericardium had been traversed and that the foreign body had entered the left side of the heart,


FIGURE 183 (case 8).-Course of migratory bullet from entry into inferior vena cava through right auricle and ventricle and out into left pulmonary artery: Apparent mechanism of entry into venous circuit (a) and course as embolus (b).

whence it passed into the arterial circulation. The case thus represents a penetrating wound of the heart or pulmonary artery with arterial embolism of the foreign body to the innominate artery.

In one case (case 8, fig. 183), the exact portal of entry of the bullet into the circulation can only be surmised. Since, however, the wound of entry was in the lower axilla and the chest was clear in the films taken immediately after wounding, it seems reasonable to assume that the bullet was originally in the liver. It probably began its migration in the inferior vena cava, whence it was carried to the right auricle, then to the left ventricle, and then into the pulmonary artery. It entered the left branch of the pulmonary artery and lodged in the branch to the lower lobe and lingula in such a way as to occlude the arterial supply to the lingula as it arose from the artery of the lower lobe in this region. Occlusion of the pulmonary circuit was associated with parenchymal changes.

No instances of intravenous migration of foreign bodies were observed at the 160th General Hospital thoracic surgery center during the war. In 9 of the 42 cases recorded in the literature, the object entered one of the larger veins, migrated to the right auricle, and came to rest in the right ventricle.


In three instances, it followed the same course as far as the right auricle, then traveled down the vena cava before coming to rest in one of its tributaries. In two instances, it entered one of the pulmonary veins and passed into the left ventricle. In other cases, the bullet entered the inferior vena cava and migrated to the junction of the common iliac veins; entered the right ventricle and passed into the pulmonary artery; entered the long dural sinus and then migrated into the sigmoid sinus; and entered the venous circulation through a wound in the thigh and was found at autopsy in the left ventricle. In the latter case, there was a widely patent foramen ovale. In the case reported by Straus (2), the bullet entered the right common iliac vein and traveled to the left pulmonary artery. Death occurred 4 days later but was not caused by the embolic missile.

In one case, a patient who harbored a missile that had migrated into the right ventricle complained of pain over the heart. With this single exception, foreign bodies migrating within the veins caused no symptoms in the recorded cases.

When a missile is in the arterial circulation, clinical manifestations vary according to the portion of the arterial tree involved. The embolic character of the object is supported by the changing symptomatology as well as by roentgenologic studies and by the findings at operation.

Varying degrees of arterial insufficiency have been reported as the result of missiles embolic in the arteries. In the case reported by Paltauf (3) and already mentioned (p. 388), death occurred when the migratory object became embolic in the right pulmonary artery. In a case reported by O'Neill (4), gangrene of the left lower extremity developed after a shell fragment became embolic from the left ventricle to the left common iliac artery; the patient survived 5 days. In one of the cases observed at the 160th General Hospital, the patient experienced severe pain when, presumably, the missile moved from the venous to the arterial circulation, and lodgment of the bullet in the branch of the pulmonary artery supplying the left lower lobe resulted in gangrene of the lingular portion of the lung on that side.

Infection-Infection is a theoretical possibility in all cases of migratory foreign bodies. Lyle (5) seems to have been the first to call attention to this fact. In his case, the foreign body, which was found at autopsy in the right ventricle, had fragments of clothing adherent to it. A fibrinopurulent pericarditis was present, and gas bubbles were noted in the myocardium, but the wound of entrance in the thigh showed no evidence of infection at any time.

Diagnosis-The possibility of intravascular migration of missiles, as the cases observed at the 160th General Hospital indicate, should be borne in mind in all instances of penetrating wounds, and the surgeon should make every effort to locate a missile whose position is not immediately obvious. If a patient who harbors a foreign body develops sudden and unexplained symptoms, as in one of these cases (case 8, p. 386), fluoroscopic and roentgenologic examinations


should be resorted to immediately, to determine whether migration of the missile is responsible for the clinical manifestations.


A review of the 33 cases collected by Straus (2), including his own, does not encourage an optimistic attitude concerning migratory foreign bodies. He listed only six survivals, and the pathologist, not the surgeon, was usually the one to remove the object and establish the course of the migration. In World War II, in view of the progress which had been made in thoracic and vascular surgery and in anesthesia between the wars, the thoracic blood vessels and the heart had come within the province of the surgeon, and there were few intravascular foreign bodies which could not be attacked surgically, and with safety. As already stated, there were no deaths in the 78 operations for the removal of intravascular foreign bodies performed at the chest center at the 160th General Hospital.

The three cases of migratory intravascular foreign bodies observed there and just described in detail, together with the cases recorded in the literature, make clear the indications for surgery, which may be stated as follows:

1. Vascular occlusion.-This risk is well demonstrated in the gangrene which ensued in the lingular portion of the left upper lobe in case 8. The pathologic process is interesting; theoretically, gangrene should not have occurred with a normal bronchial circulation.

2. Erosion and hemorrhage.-These did not occur in this series.

3. Infection.-This did not occur in any of the three cases described, but, as Lyle's (5) report indicates, it is always a possibility.

4. Embolism.-This is the overriding reason for surgical removal of the object, for a fatality is always possible, as a majority of the reported cases indicate, whenever a foreign body enters the blood vessels.

Four technical points proved useful in the management of the cases of migratory intravascular foreign bodies observed at the 160th General Hospital, as well as in a number of other operations on the mediastinum:

1. Splitting of the sternum (case 7) provided good surgical exposure. The bone was carefully approximated with wire at the end of the operation, and the patient did not suffer postoperative pain or shock.

2. Manipulation of the missile from the main right pulmonary artery to the periphery (case 6) permitted surgical manipulations in the relatively safe peripheral zone rather than in the more hazardous hilar region of the pulmonary artery. This technique, which apparently had not previously been described, was certainly less formidable than opening the right horn of the pulmonary artery, removing the missile, and then attempting local reconstitution of the main branch.

3. The digital method of hemostasis, which amounted to using the fingers as a tourniquet (case 8), proved both simple and functional.


4. As soon as the missile was encountered, it was fixed in position, so that migration could not occur during the operative procedure. The patient was also so positioned as to make circumstances unfavorable for migration. The importance of this precaution is evident from the experience in one of these cases (case 6).

One other technical point might be made concerning the case (case 7) in which incision of the innominate artery was necessary. This sort of intervention is always undertaken with some misgivings, since circulatory disturbances in the homolateral arm, as well as contralateral hemiplegia, are possible complications. These accidents, of course, were less to be feared in the age group with which military surgeons had to deal. Nonetheless, their possible occurrence points to the wisdom of restoration of vascular continuity of the innominate artery, rather than ligation, when it must be attacked surgically as in this case.


Analysis of Data

To complete the picture, the World War II experience of another chest center, at the 155th General Hospital, near Malvern Wells, England, European theater, is briefly summarized. It included the removal of 172 objects from the lungs; 27 from the mediastinum; 37 from the pleura; and 19 from various internal structures, in addition to 63 removed from the deep structures of the thoracic wall. The 172 cases classified as pulmonary included foreign bodies situated in the paramediastinal surface of the lung, either abutting on the mediastinum or partly embedded in the mediastinal pleura.

In four other operations, the foreign body was not removed. In two instances, it lay within the substance of the lung at the hilus but escaped detection because of hematoma formation and pulmonary infiltration. In both of these cases, the object was later removed without difficulty. In the two remaining cases, the object was displaced at operation. In one instance, it lay within the pulmonary artery on one side and slipped into the artery on the opposite side during manipulations. In the other case, it lay in the right ventricle and was displaced into the right auricle and inferior vena cava.

About 93 percent of the retained foreign bodies were high explosive shell fragments. The remaining injuries were caused by small arms fire.

Of the 172 foreign bodies in the lung, 119 were removed by transpleural pneumonotomy. In 48 cases, the free pleura was not entered; the operative procedure could be carried out entirely within a zone of adhesions between the lung and the thoracic parietes. In the remaining five cases, in which a collection of pus was encountered unexpectedly around the object at operation, removal was transpleural, and local marsupialization of the lung was per-


formed in order to minimize the danger of pleural infection. This policy, as already intimated, was successful.

The incision, whenever practical, was placed directly over the foreign body and was always as small as was consistent with satisfactory surgical exposure. A rib resection was never performed when intercostal incision and division of the rib were adequate. When the object could be removed through adhesions which existed between the lung and the pleura, postoperative pleural effusion did not occur, and convalescence was appreciably shortened. The technique employed in such cases was similar to that used in one-stage drainage of an abscess of the lung. Particularly small incisions were required when it was employed.

Of the 27 foreign bodies in the mediastinum, 7 lay in direct relationship with the esophagus; 7 with the pericardium; 6 with the aorta; 2 each with the azygos vein and the innominate artery; and 3 with the vena cava. Seventeen of the objects were removed transpleurally and ten by extrapleural mediastinotomy, the latter procedure being used exclusively when the foreign body lay in the anterior or posterior mediastinum.

The 37 foreign bodies classified as pleural either lay free in the pleural cavity or were embedded superficially in the pulmonary diaphragmatic or parietal pleura and projected into the cavity. In many of these cases, removal was possible through an intercostal incision, without the necessity for rib resection or rib division.

The foreign bodies removed from the deep structures of the thoracic wall involved the peripleural region, the ribs, the intercostal muscles, and the root of the neck immediately adjacent to the dome of the pleura. Except for four cases in which the free pleura was inadvertently entered, all of these objects were approached through small, accurately placed incisions, for which careful preoperative roentgenologic studies had served as guide.

In the remaining 19 cases, 4 foreign bodies were removed from the diaphragm, in which they were completely embedded; 3 each from the liver and the pericardium; 2 each from the pulmonary artery and the heart; 1 from the subphrenic space; and 4 from the vertebral column.

Postoperative management followed standard practices. At first, penicillin was used intrapleurally as well as parenterally. Later, only the parenteral route was employed, since the intrapleural route was thought to increase the frequency and size of postoperative effusion.

There were no pleural infections in the 318 operations in this series, 255 of which were for the removal of intrathoracic foreign bodies. The single death was difficult to explain: A large foreign body, situated posteriorly and superficially in the lung, was removed through a zone of visceroparietal adhesions, without entry into the free pleura. Eight hours after operation, while the patient was lying in bed watching a moving picture being shown in the ward, he suddenly lost consciousness and developed jacksonian seizures involving the right arm and right leg. Death occurred 16 hours later. At autopsy, a cere-


bral hemorrhage was found, the precise etiology of which could not be determined.


In general, indications for the removal of retained foreign bodies at the 155th General Hospital were conservative, a combination of size (not less than 1 cm. in diameter) and symptoms, of which pain was most prominent. Objects less than 1 cm. in diameter were removed only if they were jagged or irregular or if they were located close to some vital structure and there appeared to be danger of erosion because of mechanical factors or infection.

As the war progressed and the maximum permissible period of hospitalization was progressively reduced, certain patients who would previously have been operated on overseas were returned to the Zone of Interior for surgery. Such cases included those in which the foreign bodies were in, or adjacent to, the heart and great vessels, as well as those in the upper abdomen and in, or adjacent to, the liver. Since most patients with intrapulmonary and intrapleural foreign bodies could be returned to duty overseas within a relatively short time, the proportion of operations in these groups remained fairly stationary throughout the war. Otherwise, operation was not done overseas unless the existence of suppuration or some other complication made the procedure urgent.


1. Turner, G. Grey: A Bullet in the Heart for Twenty-three Years. Surgery 9: 832-852, 1941.

2. Straus, R.: Pulmonary Embolism Caused by a Lead Bullet Following a Gunshot Wound of the Abdomen. Arch. Path. 33: 63-68, January 1942.

3. Paltauf, R.: Geschossembolie der Arteria pulmonalis. Wien. klin. Wchnschr. 46: 602-603, 1933.

4. O'Neill, C. S.: Fragment of Shell in the Arterial Circulation. Brit. M.J. 2: 719-720, 1 Dec. 1917.

5. Lyle, H. H. M.: Migration of Shell Fragment From Right Femoral Vein to Right Ventricle of Heart. Generalized Gas Bacillus Infection. J.A.M.A. 68: 539, 17 Feb. 1917.