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 III



Thoracoabdominal Wounds

Lyman A. Brewer III, M.D.


Definition-The term "thoracoabdominal wound" was reserved in World War II for a wound in which a single missile was responsible for damage to structures of the chest and the abdomen. In approximately 90 percent of all cases, the missile entered the chest and passed into the abdomen. As a rule, the term "thoracoabdominal injury" implied a perforation of the diaphragm. Occasionally, a missile that entered the chest could transmit sufficient force through all an intact diaphragm to produce an abdominal injury, but this was most unusual. A wound of the so-called bare area of the liver produced by a fragment that entered the chest and perforated the diaphragm was considered a true thoracoabdominal wound.

The term "thoracoabdominal wound" should never have been used, as it occasionally was, for wounds in which both serous cavities were injured by separate missiles. The distinction is important. The management of a combined thoracic and abdominal wound inflicted by separate missiles and without perforation of the diaphragm differed considerably from the management of a thoracoabdominal wound in which the injury of both cavities was produced by the same missile and in which perforation of the diaphragm was a factor. Thoracoabdominal and combined thoracic and abdominal wounds resembled each other in one respect, however, that both carried very high mortality rates.

Incidence-There was not complete agreement as to the incidence of these wounds.1 They accounted for about 5 percent of all admissions for battle-incurred wounds in Seventh U.S. Army hospitals. Maj. (later Lt. Col.) Lawrence M. Shefts, MC, and Capt. (later Maj.) Ernest A. Doud, MC, who had a representative forward experience, found that thoracoabdominal wounds represented about a quarter of their intrathoracic wounds. The 903 thoracoabdominal injuries treated by the 2d Auxiliary Surgical Group amounted to almost 40 percent of the 2,267 intrathoracic injuries which they encountered.

Management in field hospitals-About a quarter of the battle-incurred injuries treated in field hospitals in the Mediterranean theater were thoracoabdominal injuries. Since casualties with these injuries were always nontransportable, practically all were treated in these forward hospitals. The

1Official statistics for thoracoabdominal injuries appear in the first volume dealing with thoracic injuries (chapter II, tables 8-11).


occasional casualty treated in an evacuation hospital had an injury that had escaped recognition in the field hospital. The management of thoracoabdominal injuries in base hospitals consisted of the management of complications and the care and closure of colostomies.

As a group, casualties with thoracoabdominal wounds presented the severest injuries of all battle casualties; furnished a high proportion of the admissions to field hospitals, where, as just noted, most of them were treated; and had the highest case fatality rate of any group admitted to them. 

These wounds furnished the most important single indication for surgery of thoracic injuries in a field hospital, for two reasons, (1) that resuscitation was not complete without surgical repair of the intraperitoneal damage, and (2) that in this type of wound, immediate surgery was necessary to prevent or control intraperitoneal contamination. In thoracoabdominal wounds, whether the diagnosis was established or merely probable, surgery was indicated as soon as cardiorespiratory stability permitted.

Attention has been called elsewhere to the necessity for thoracic surgeons assigned to field hospitals to be able to do intra-abdominal surgery (vol. I). The large number of thoracoabdominal wounds that had to be handled in these installations was further proof of the need for this dual ability. Even though the majority of casualties with injuries limited to the thorax could be evacuated to the evacuation hospitals after resuscitation, if not before, a single team could not handle the remainder of the casualties in this group plus all of the casualties in the thoracoabdominal group. Thoracoabdominal injuries were practically always very serious, and a great deal of time was consumed in resuscitation and at operation. A desirable plan, when available personnel permitted its use, was to assign a senior thoracic surgeon to a field hospital and place two teams under his direction. In this way, he could supervise the management of all serious thoracic and thoracoabdominal wounds without attempting the usually impossible task of managing them all personally.


The wound of entrance in thoracoabdominal wounds was in the thorax in the great majority of cases. In the 903 such wounds encountered by the 2d Auxiliary Surgical Group, the missile entered through the chest 837 times and through the abdomen 66 times.

Thoracic injuries-In most instances, the entering missiles caused fractures of the ribs; the level of the fractures tended to correspond with the zone of the diaphragm. In only 3 of the 903 wounds just mentioned was the site of the costal fracture above the sixth rib posteriorly or the fourth rib anteriorly. The paucity of wounds in the upper chest is probably to be explained by the smaller size of the target and by the lethal effect of missiles which entered the thorax in a cephalad direction; these patients did not survive to reach a hospital.


Pulmonary damage varied according to the site of injury. When a costophrenic angle was crossed, damage was usually minimal. If the wound of entrance was elsewhere, there was usually a perforation or a laceration of the lung. Contusions were less frequent. Hemorrhage was present along the track of the missile and for 1 to 3 cm. beyond it. In an occasional case, a segment of the lung was found amputated.

The mediastinum was injured 30 times, the heart 14 times, and the pericardium alone 13 times in these 903 cases. There was only one injury of the esophagus.

In most cases handled by the 2d Auxiliary Surgical Group, on whose material all of the following statements are based, the pulmonary wound was in the peripheral portions of the lower lobes because the missiles most often entered through the lower chest in the zone of the diaphragm. Injuries of the middle lobe and of the lingula of the left upper lobe were much less frequent.

Diaphragmatic injuries-Injuries to the diaphragm (fig. 35) fell roughly into three categories:

1. Small single or double perforating wounds.

2. Large lacerated wounds.

3. Avulsion of the diaphragm from its attachment to the chest wall.

The wounds were almost equally distributed between the right (435 cases) and the left (448 cases) sides. In 20 cases, the diaphragm was perforated bilaterally.

Herniation of abdominal contents through the diaphragm was recorded 57 times. In 17 instances, only the omentum protruded. The explanation of evisceration was the physiologic fact that pressure is normally lower within the chest than within the abdomen. As a result, the spillage of gastric and intestinal contents into the chest was always a possibility when the diaphragm had been wounded.

Intra-abdominal injuries-The liver took the impact of missiles which entered on the right side. It was involved in 407 of the 435 right-sided injuries. In the order of frequency after wounds of the liver (table 7) were wounds of the right kidney, colon, stomach, small bowel, and gallbladder. On the left side, the spleen, stomach, and colon took the impact of most missiles entering on that side. In the 448 left-sided wounds, the spleen was injured 272 times, the stomach 167 times, and the colon 145 times.

There was no evidence of involvement of the thoracic duct in these 903 cases. The assumption is that casualties with such injuries seldom survived long enough for their condition to be recognized.

Tension pneumothorax of peritoneal origin-In the 55 thoracoabdominal wounds encountered by Lt. Col. John M. Snyder, MC, and Maj. Frank Tropea, Jr., MC, 24 of which were left-sided, there were 2 instances of tension pneumothorax of peritoneal origin. This condition, it was thought, had not previously been described. In each instance, the wound of entry was in the lower left chest, and in each, there was a large perforation of the anterior wall


FIGURE 35.-Types of wounds of diaphragm possible from penetrating or perforating missiles. A. Through-and-through wound of lower chest. On inspiration, the wound may be purely thoracic. B. Same type of wound, which may be thoracoabdominal on expiration. C. Thoracoabdominal wound with, as usual, missle entering chest first. D. Thoracoabdominal wound with missile entering abdomen first.

of the stomach, with a resultant outpouring of a large amount of air through the lacerated diaphragm. In the first case, the diagnosis was not made, and death occurred in the course of laparotomy. In the second, in which the condition was recognized several hours after the patient was first seen, closed drainage was instituted, and recovery was uneventful.

Theoretically, there seems to be no reason why tension pneumothorax of peritoneal origin should not be possible under the circumstances described,


TABLE 7.-Frequency of wounds and case fatality rate in 903 thoracoabdominal wounds according to combinations of viscera

Combination of organs



Case fatality rate

Liver alone




Liver and kidney




Liver and stomach




Liver and small intestine




Liver and spleen




Liver and right colon




Liver and left colon




Liver, stomach, and colon




Liver, kidney, and right colon




Liver, stomach, and biliary tract




Liver, stomach, and spleen




Liver, small intestine, and left colon




Stomach and spleen




Stomach, spleen, and left colon




Stomach, spleen, and kidney




Stomach, spleen, and pancreas




Stomach and left colon




Stomach, small intestine, and colon




Stomach and kidney




Spleen alone




Spleen and kidney




Spleen and left colon




Spleen, kidney, and left colon




Spleen and small intestine




Kidney alone




Left colon alone




Left colon and small intestine




Peritoneal cavity (penetration only)




All other combinations1




When five organs or more were injured, the case fatality rate was 100 percent.

though before autopsy established it in the fatal case, some doubt had been felt about it. A tension pneumothorax of peritoneal origin could scarcely occur in a thoracoabdominal wound on the right, but the possibility of what might be termed an internal sucking wound could reasonably be considered on the left side. In this variety of wound, small quantities of air and intestinal contents could be aspirated into the chest during the act of respiration, and the repeated aspiration of small quantities would alter the intrapleural pressure and lead to pulmonary collapse unless the sucking was promptly corrected.

During the laparotomy for a left-sided thoracoabdominal wound, an injury that originated as an internal sucking wound would become an external sucking wound as soon as the peritoneum was opened. This is apparently what happened in the fatal case in this series. Next to the control of active hemorrhage,


therefore, closure of the lacerated diaphragm was the first duty of the surgeon who operated on a thoracoabdominal wound through a laparotomy incision. On the right, the liver usually sealed off the diaphragmatic wound. 

In the early days of the war, many laparotomies were performed without endotracheal anesthesia, and as a result, when the abdomen was opened, a tension pneumothorax developed through the perforation in the diaphragm. This complication ceased to be a problem after endotracheal anesthesia was in general use and surgeons became aware of the possibilities of tension pneumothorax of this origin.


In general, if the patient had a single wound of the thorax and a single wound of the abdomen, the diagnosis was obviously a combined thoracic-abdominal wound. If both wounds were in the abdomen, there was little possibility of diaphragmatic injury unless the costal cage was involved. If both wounds were thoracic and if one was below the level of the fourth rib anteriorly, a thoracoabdominal injury might or might not be present.

Since not more than 10 percent of the missiles in thoracoabdominal wounds entered the body via the abdomen, the chief difficulty from the diagnostic standpoint was in determining the presence or absence of abdominal injury in a patient with a known thoracic wound, in every one of which penetration of the abdomen had to be considered as a possibility until the injury was excluded. There was no problem when roentgenologic examination showed the retained foreign body to lie within the chest or showed that it had entered the abdomen. The diagnostic difficulties were posed by the missiles that perforated the chest in the so-called diaphragmatic zone or penetrated it and lodged in the vicinity of the diaphragm.

Anatomic and Physiologic Considerations

The most practical method of diagnosing a thoracoabdominal wound was to keep constantly in mind that the diaphragm might be involved in any thoracic injury, most particularly in any injury occurring in the area bounded superiorly by the anterior end of the fourth rib and posteriorly by the level of the seventh rib, down to and including the twelfth rib at its inferior level (fig. 36).

Both physiologic and anatomic facts had to be borne in mind in making the diagnosis. It could be assumed that the pleural reflection follows, with reasonable accuracy, the outline of the periphery of the ribs and the costal cartilages. For diagnostic purposes, the thickness of the diaphragm could be ignored, and the peritoneum could be assumed to be in contact with the diaphragmatic surfaces of the pleura. The diaphragm, however, is not an immobile structure which forms a perfectly horizontal partition between the chest and


FIGURE 36.-Schematic showing of maximum expiratory excursion of diaphragm. In deep expiration (a), it can reach as high as the superior border of the fourth rib anteriorly. Note depth of costophrenic sulcus on inspiration (b) draws the diaphragm down to the sixth rib.

the abdomen. In these circumstances, diagnosis would be much simpler. Instead, the variables caused by its uneven configuration and the distance through which it can move, tremendously increased its susceptibility to injury, and, in turn, the susceptibility of abdominal organs to involvement in the wound.

Because of the movement of the diaphragm, any wound of the entire lower left hemithorax might penetrate the division between the thorax and the abdomen (fig. 36). Any wound below the seventh interspace posteriorly or the fourth rib anteriorly might injure the diaphragm if the patient was in the expiratory phase of respiration at the time. Since the pleural reflection extends down to the attachments of the diaphragm to the ribs and the costal arch, any wound at the level of the twelfth rib or above posteriorly, or any wound that involved the costal arch or above anteriorly, had to be considered a possible thoracoabdominal injury.

If the range of motion of the diaphragm on deep inspiration and expiration were kept in mind, the cavities and organs involved in a wound of the chest might reasonably be predicted by lining up the wounds of entrance and exit if the injury were perforating, or by lining up the wound of entrance with the point at which the missile was lodged within the body if the injury were penetrating. Any missile, therefore, which passed through the area extending from the level of the twelfth rib posteriorly to the level of the fourth rib anteriorly might or might not have entered both the pleural and the peritoneal cavities.


Since many wounds were inflicted while the soldiers were prone, the increase in intra-abdominal pressure created by this position further raised the upper limits of the diaphragm.

If the patient was in deep expiration at the time of wounding, the diaphragm, with the spleen and stomach, might rise high enough into the thoracic cavity to be injured by a missile that had entered the chest at the fourth interspace. If the wounding had occurred during the expiratory phase of respiration, the anatomic position of the wounds in the chest wall might be quite cephalad and yet there might be a tear in the diaphragm and damage to the peritoneal organs. In one casualty, for instance, in whom the wound of entrance was in the fourth interspace just medial to the left nipple and the wound of exit was in the seventh interspace in the left midaxillary line, the diaphragm was perforated and the spleen was injured.

Projection of Course of Missile

The most reliable and most helpful information in thoracic injuries in which involvement of the abdomen was not evident by clinical means was obtained by projecting the probable course of the missile. Roentgenologic examination was therefore essential in every penetrating wound of the chest, no matter how insignificant it might seem, if only because the location of the wound of entrance was occasionally extremely misleading in relation to the course of the missile within the body.

Both anteroposterior and lateral films were made, with overexposure for both abdomen and chest. They were preferably made in the sitting position. Patients in severe shock, who could not tolerate the upright position, were examined by anteroposterior exposures made in the prone position.

When the information derived from films in two planes was correlated with its point of entrance, the course of the missile could be projected and a decision made concerning the possibility of intra-abdominal damage.

Missiles tend to travel in a straight line if they are not deflected. The experience of most surgeons who treated large numbers of war wounds was, as pointed out elsewhere (vol. I), that stories of missiles which entered at one point and caromed around at a tangent before making their exit or lodging at some point within the body were the exception. There were occasional instances in which the course of the missile was bizarre because it had struck a bony structure tangentially, but these injuries, again, were exceptional. Many superficially bizarre pathways could also be explained by investigation of the position that the casualty was occupying when he was injured. Inquiry on this point was made whenever the patient was sufficiently oriented to furnish the information.

On the other hand, an occasional thoracoabdominal injury diagnosed for the first time at a general hospital made it clear that exceptions to these generalizations sometimes occurred. In one such injury, reported by Lt. Col. John Burke, MC, and Maj. Theodore T. Jacobs, MC, for instance, a wound of


entrance in the region of the right scapula and a wound of exit in the right groin were considered to be individual wounds. They proved to be from the same missile. Fortunately, the delay in operation was not attended with serious consequences.

Clinical Considerations

Part of the diagnostic confusion in some thoracoabdominal wounds was caused by the overshadowing of symptoms and signs referable to the abdominal injury by the more urgent symptoms and signs of the chest injury. The thoracic component of the injury could give rise to pain in the chest wall, hemoptysis, hemothorax or hemopneumothorax, dyspnea, cyanosis, and wet lung.

Since thoracic wounds that did not involve the abdomen were frequently associated with pain and spasm of the upper abdomen (vol. I), these phenomena did not furnish a reliable indication of abdominal injury. Abdominal tenderness and rigidity were also often present; in such cases, a useful differential point was that deep tenderness and rebound tenderness were never elicited in chest wounds. If time could be spared, a differential diagnosis was sometimes clarified by paravertebral or intercostal block. The results were not always conclusive, but when there was prompt relief of pain, abdominal involvement could usually be excluded. Pain in the shoulder was not a universal symptom. When it was present, it pointed to an injury of the diaphragm.

Although nausea and vomiting sometimes occurred in thoracic injuries, they were not common, and their presence lent strength to the possibility of abdominal involvement. The demonstration of active peristalsis suggested that if abdominal injury had occurred, a hollow viscus was not involved. Most surgeons, however, could recall casualties in whom peristalsis had been observed even with wounds of the colon. Peristalsis frequently persisted when the injury was limited to the spleen, liver, or kidney. Wounds of all of these organs were associated with a paucity of symptoms and signs referable to the abdomen.

If a hollow viscus was injured, prompt surgery was imperative, but it was important to remember that injuries of solid viscera could also be extremely serious. A casualty might bleed to death from a damaged spleen, or grave or even fatal bile empyema might be the consequence of an overlooked wound of the liver.

Other Diagnostic Aids

The presence of blood in gastric contents, whether vomited or aspirated, was suggestive of abdominal injury, though it was not conclusive, since it might have originated from a wound of the lung.

Examination of a voided or catheterized specimen of urine was part of the diagnostic routine. The absence of blood in the urine was not conclusive, but its presence furnished valuable information as to the course of the missile in relation to the urinary tract.


Such diagnostic refinements as pneumoperitoneum were not practical in forward hospitals.



The final diagnosis of a thoracoabdominal wound was based on the following considerations:

1. Inspection of the wound or wounds.

2. The projection of the suspected course of the missile from the wounds of entrance and exit if the injury were perforating or from roentgenologic localization of its position if the injury were penetrating.

3. Such information as could be obtained from the history and from the general clinical picture.

If a positive diagnosis of a thoracoabdominal wound could not be made with reasonable promptness and the suspicion persisted that the diaphragm might have been perforated, prompt exploration was the rule. A positive preoperative diagnosis was naturally desirable, but if it could not be made, a reasonable suspicion that such an injury existed was sufficient indication for exploration. In many instances, a somewhat enlarged traumatic thoracotomy provided adequate exposure to determine whether or not the diaphragm had been perforated.

The only suspected thoracoabdominal injuries in which exploration was not regarded as mandatory were minor penetrations of the liver. If a foreign body was only 2 or 3 mm. in diameter, and if roentgenologic examination made it unmistakably clear that it lay entirely within the liver, expectant treatment was considered permissible in certain selected cases. Not more than 1 percent of all thoracoabdominal injuries fell into this category, and many surgeons considered the risk of biliary empyema too considerable to warrant expectant treatment in any case at all. Whenever the foreign body was over 3 mm. in diameter, regardless of its location, it was always best to determine surgically the exact damage that had occurred.

Negative Explorations

A rather large number of exploratory thoracotomies had to be performed because of the difficulty of making an accurate diagnosis of wounds around the diaphragmatic sulci. For this and other reasons, a significant number of negative explorations were carried out in all field hospitals. In the 2,267 penetrating and perforating wounds of the chest cared for by thoracic surgeons of the 2d Auxiliary Surgical Group, there were 903 proved thoracoabdominal injuries, almost 40 percent. Teams from this group, however, performed 122 exploratory thoracotomies in which no abdominal involvement had occurred.


These 122 negative explorations represent 28 percent of all thoracotomies done by these teams in forward hospitals.

To surgeons without experience in combat wounds, this figure may appear high. It is certainly high compared with the 41 explorations (1.3 percent) performed by surgical teams of the group in 3,154 abdominal wounds in which no intraperitoneal damage of any consequence was found (1). On the other hand, it is significant that well over half of these 41 negative explorations (28) were in thoracoabdominal wounds. In these 28 explorations, as in the 122 negative explorations performed for suspected thoracoabdominal wounds in this series of thoracic injuries, the indication was the same: Injuries of the abdominal viscera are so highly lethal if they are not treated surgically that no chance of overlooking them could be taken.

A practical consideration in the diagnosis of thoracoabdominal wounds was the echelon of medical care in which surgery was done. Casualties with thoracoabdominal wounds in which the diagnosis was clear cut were not transportable; surgery was done in field hospitals. Casualties in whom thoracoabdominal wounds were suspected also had to be cared for in field hospitals. Exploration could not be delayed until they reached an evacuation hospital.2 If an abdominal injury were positively excluded, surgery for the chest injury could usually be deferred until an evacuation hospital was reached. Unnecessary exploration put an added strain on the space, facilities, and personnel of the field hospital, but it could not be omitted in any doubtful case.


General Considerations

Certain concessions to the realities of the situation had to be made in the resuscitation and preoperative preparation of casualties with thoracoabdominal injuries. The exigencies of the abdominal emergency were sometimes met at the expense of the most ideal preparation of the casualty from the point of view of cardiorespiratory stabilization.

From the standpoint of pure thoracic injuries, the length of time required, within reasonable limits, to restore thoracic casualties to as nearly normal status as possible before operation was not of primary importance. If the injury was limited to the thorax, the casualty frequently was benefited by a short period of stabilization after cardiorespiratory abnormalities had been corrected and shock controlled.

In casualties with thoracoabdominal wounds, the situation was entirely different. Time was of great importance. They had to be resuscitated as expeditiously as possible and operated on as soon as it was thought that they

2In certain military situations, evacuation hospitals served as the most forward hospitals in the area and acted in the capacity of field hospitals. In these circumstances, they operated on all nontransportable casualties.


could tolerate surgery. If a patient continued to have a low blood pressure, rapid pulse, and associated signs of shock after correction of the disordered thoracic physiology, supplemented by adequate replacement therapy, immediate surgery was indicated. Many patients who presented such a picture were found at operation to be bleeding seriously or to have massive pleural or peritoneal contamination from a perforated stomach or intestine. Their condition was frankly dangerous with immediate surgery, but entirely hopeless without it.

Patients with thoracoabdominal injuries were examined in the shock tent to determine their status in respect to shock and their cardiorespiratory status as manifested by respiratory restriction due to pain, embarrassment of respiration by hemothorax or pneumothorax, cyanosis, and excessive accumulations of blood and mucus in the tracheobronchial tree. In 669 thoracoabdominal injuries in the 2d Auxiliary Surgical Group records in which information is available on this point, 145 patients were not in shock but 75 were in mild shock, 174 were in moderate shock, and 275 were in severe shock.

Routine of Resuscitation

When thoracoabdominal casualties were received in the field hospital from the division clearing station, measures to combat shock and maintain respiratory equilibrium would already have been instituted. Sucking wounds of the chest would have been closed by occlusive dressings. Pain would have been relieved by morphine if it was indicated, or by intercostal nerve block. Plasma transfusions would also have been given as necessary. The 2d Auxiliary Surgical Group records in which information is available on this point show that 144 casualties had received no plasma, but 455 had received an average of 525 cc. each.

The records also indicate that while 60 casualties received no blood in the shock tent, 619 had an average of 1,100 cc. each. It is known that 252 casualties received no plasma in the shock tent and that 408 received an average of 500 cc. each. The usual practice was to begin the administration of plasma in the shock tent while waiting for blood to be cross-matched.

While plasma and blood were being administered, measures were instituted to correct the disordered cardiorespiratory physiology, including the administration of oxygen through a nasal catheter or mask (vol. I); intercostal nerve block if it had not already been carried out (vol. I); aspiration of air and blood from the pleural cavity (vol. I); and catheter or bronchoscopic aspiration of the tracheobronchial tree (p. 227). As a rule, the effect of these measures upon the cardiorespiratory imbalance and on the state of shock to which it was contributing was most gratifying.

As pointed out already (vol. I), a patient with a thoracic injury had to be treated judiciously in respect to replacement therapy because of the grave danger of overloading the already damaged cardiorespiratory system and pre-


cipitating pulmonary edema. The same precautions had to be observed in thoracoabdominal injuries. However much the casualty might need blood from the standpoint of the abdominal component of his wound, there was serious risk from the standpoint of the thoracic component in giving large amounts acid in giving them too rapidly.

An important step in preoperative management was the introduction of a Levin tube, to aspirate the gastric contents and thus eliminate the risk of vomiting during surgery (vol. I).

It was a general policy to transport the patient from the shock tent to the operating table on the litter on which he had been placed originally, to operate on him on the same litter, and to move him to the recovery tent on it. Casualties with abdominal and thoracoabdominal wounds, particularly those who had been in serious shock, did not tolerate movement well, and this plan eliminated it while at the same time saving the time and effort of busy personnel.

Priority of Surgery

Uness severe intro-abdominal hemorrhage, which was not usual, demanded priority of treatment, the necessary intrathoracic surgery took precedence of the abdominal surgery. The importance of this sequence is evident in 15 deaths analyzed by Maj. (later Col.) Howard E. Snyder, MC, Consultant in Surgery, Fifth U.S. Army (2). In only one of seven cases in which laparotomy preceded chest surgery was there any indication for performing it first, and there seemed no doubt that the unwise order of precedence played a part in more than one of the fatalities.


The anesthesia of choice in thoracoabdominal wounds was endotracheal ether and oxygen in a closed circuit, with induction by nitrous oxide and oxygen. In a small number of cases, Pentothal sodium (thiopental sodium) was used for this purpose. In 1942 and part of 1943, open drop ether was used in many operations because of the limited number of anesthetic machines. This technique was discarded as soon as these machines became generally available.

It was impossible to overemphasize the value of endotracheal anesthesia, in thoracoabdominal surgery or of the services of an expert anesthesiologist. Such an anesthesiologist appreciated the valve of frequent and thorough aspiration per catheter of the tracheobronchial tree throughout the operation; of bronchoscopy in selected cases at the end of operation; of replacement therapy during operation; and of positive pressure anesthesia when the thoracic cavity was open as well as during closure of the chest, when reexpansion of the lung was necessary.


Replacement therapy was carried out during surgery, the use of whole blood, plasma, or glucose or physiologic salt solution depending upon the needs of the individual patient. The records of the 2d Auxiliary Surgical Group that are complete in this respect show that 116 patients required no fluids on the operating table, 639 had an average of 1,000 cc. each of blood, 166 an average of 575 cc. each of plasma, and 78 an average of 1,000 cc. each of 5 percent glucose in physiologic salt solution in addition to the plasma and blood they received.

The casualties operated on by teams of the 2d Auxiliary Surgical Group were kept under anesthesia on an average of 2 hours.


General Considerations

At the beginning of the fighting in North Africa, it was generally thought that the transdiaphragmatic approach to the abdominal component of thoracoabdominal wounds would be of only limited usefulness. At that time, and to a certain extent throughout the war, the decision as to the surgical approach depended not only upon the findings in the particular case but also upon the preference of the surgeon for the abdominal or thoracic approach. This preference usually depended, in turn, upon his experience, or lack of experience, in chest surgery. The experienced chest surgeon was able to select the single approach or the combined approach which best met the needs of the patient. Surgeons not widely versed in thoracic surgery found it safer to repair the intraperitoneal damage through a laparotomy incision after handling the thoracic damage and closing the diaphragm through the chest.

By the end of the war, it was the opinion in the Mediterranean theater-an opinion shared by the more experienced of the thoracic surgeons in the European theater-that whenever it was practical, thoracotomy with transdiaphragmatic repair of all accessible abdominal injuries was infinitely the procedure of choice. If total repair was not possible by the transthoracic approach, the best plan in left-sided wounds was to correct the intrathoracic damage, care for any damage readily accessible in the left upper quadrant of the abdomen, close the diaphragm, reexpand the lung, close the chest wall tightly, and then proceed with the repair of intraperitoneal injuries not accessible through the diaphragm. The same general principle was applicable to wounds of the upper right quadrant of the abdomen, but the extent of surgery possible in this quadrant was, for anatomic reasons, much less than on the left side.

When the thoracic and abdominal wounds were separate, transdiaphragmatic surgery was not indicated because it subjected the pleura to contamination which did not already exist. In these cases, the chest wound was given priority unless profuse intra-abdominal bleeding or massive evisceration made laparotomy necessary as the primary procedure.


Advantages and Disadvantages of the Transthoracic Approach

No hard-and-fast rule, covering all casualties, could be laid down for the approach to a thoracoabdominal wound. There were advantages and disadvantages to both the thoracic and the abdominal approaches, and each casualty had to be individualized.

In favor of the transthoracic approach were the following arguments:

l. If there was extensive thoracic damage, the chest injuries could not be repaired from below the diaphragm.

2. Certain wounds of the upper abdomen were more easily handled through the diaphragm than through a laparotomy incision. These included wounds of the spleen, the dome of the liver, both kidneys, the upper portion of the stomach (especially if the posterior surface was injured), and the splenic flexure of the colon. There was real danger of overlooking wounds of the posterior wall of the stomach and of the retroperitoneal portion of the splenic flexure, as well as wounds of certain areas of the spleen, when the approach was by laparotomy.

3. Injuries of the diaphragm, whether on the right or the left side, especially if they were extensive, were more readily repaired transthoracically. If the wound was small, it could as well be repaired from below as from above. Injuries over the dome of the liver, however, as well as such large defects as resulted from tearing of the diaphragm from the costal margin, especially posteriorly, were much more readily repaired from above.

4. The transthoracic approach permitted exteriorization of the transverse or splenic flexure of the colon through a subcostal gridiron incision at a greater distance from the surgical incision than was possible if laparotomy had been done. Wound infection was thus reduced to a minimum.

5. If considerable pleural contamination had occurred from spillage of gastric or intestinal contents, it could be reduced by copious pleural lavage with physiologic salt solution, a procedure possible only through the chest.

6. There was no validity to the objection, sometimes raised, that wounds of the gastrointestinal tract should not be dealt with through the chest because infection of the pleural cavity might result. For one thing, considerable damage to the intrathoracic organs might exist without being evident before the chest was explored. For another, potential infection existed whenever the diaphragm had been perforated by the missile. If hollow organs in the abdomen had been perforated, feces and gastric contents had been sucked into the pleural cavity before the surgeon ever saw the casualty.

7. If the chest was long and narrow, with an acute angle at the costal arch, subdiaphragmatic repair was often technically very difficult.

8. Blockage of the intercostal nerves, for anatomic reasons, was much easier when the transthoracic approach was used. These nerves lie underneath the ribs and course downward diagonally to supply the upper abdomen. When they were blocked at operation, both thoracic and abdominal pain was decreased. As a result, after operation, the patient coughed willingly, and as


a further result, the lungs were more adequately aerated and bronchial secretions were raised more completely.

9. Since abdominal relaxation was not necessary when the transthoracic approach was used, the patient could be carried in a lighter plane of anesthesia.

10. If the transthoracic approach revealed intra-abdominal injuries that could not be repaired from above, it was a simple matter to perform the necessary chest surgery, close the diaphragm and chest wall, and then perform the necessary intraperitoneal surgery through a separate laparotomy incision. As already emphasized (p. 113), in the absence of compelling reasons for reversing the order, chest surgery was always done first, in order to restore normal cardiorespiratory conditions as promptly as possible.

Against these multiple advantages of the transthoracic approach were the following arguments for repair of intraperitoneal damage through a laparotomy incision:

1. Certain wounds of the small bowel and colon could not be exposed adequately through the diaphragm and had to be repaired by laparotomy. These included wounds of the terminal 18 inches of the small intestine and all wounds of the colon except wounds of the transverse portion, the splenic flexure, and the proximal portion of the descending colon.

2. When thoracic damage was minimal, an abdominal approach avoided unnecessary entrance into the pleural cavity. This was not a sound argument. When the abdomen was opened, all but the most minimal perforations of the diaphragm constituted, in effect, sucking wounds, which carried all the risks of sucking wounds of the chest wall (p. 105). Not infrequently, patients developed pulmonary embarrassment while the abdominal repair was in process because of aspiration of air into the chest through the diaphragmatic opening. If the situation was unrecognized and uncorrected, serious results could follow, including aspiration into the pleural cavity of intestinal and gastric contents free in the peritoneal cavity. If the abdominal approach was used, repair of the diaphragm was necessarily the first step of the procedure, to avoid these dangers, instead of the last before wound closure, as in the transthoracic approach.


Four types of incisions, in addition to separate thoracic and abdominal incisions, were employed in thoracoabdominal wounds:

1. Limited thoracotomy, accomplished by an extension of the original wound (traumatic thoracotomy). This was often a satisfactory incision when injuries were confined to the chest. It was not usually satisfactory in thoracoabdominal wounds unless the original injury in the chest wall was large enough to supply adequate exposure, which was essential for exploration of the upper abdomen and repair of damaged intra-abdominal viscera. If the original wound was below the tenth rib, and especially if the missile had gone through the eleventh or the twelfth rib, traumatic thoracotomy was entirely


impractical. Even resection of one or more ribs did not provide satisfactory space for maneuvers. Wide exposure was essential. If there was any reason for inspection of the interior of the thorax, there was every reason for adequate investigation.

2. Thoracotomy in the area of the original wound.

3. Thoracotomy outside the area of the original wound. The majority of thoracotomies fell into either this group or the second group. The formal approach was usually posterolateral and usually through the ninth intercostal space or the bed of the ninth rib. Injuries of the mid or anterior portions of the diaphragm were often approached through the eighth interspace or the bed of the eighth rib. A small number of thoracotomies were performed laterally, in the region of the sixth or seventh ribs, more often through the intercostal spaces than through the bed of the ribs. It was always best, if it was anatomically feasible for good exposure, to resect a rib that was not involved in the wound of entrance or exit, in order to reduce the chances of postoperative wound infection with involvement of the diaphragm.

4. Thoracolaparotomy, an incision in which the thoracic incision was extended through the costal arch onto the abdomen. This was not a popular incision because of the resulting instability of the chest wall and the increased morbidity if infection occurred in the costal cartilages. It was used only six times in the 903 thoracoabdominal wounds cared for by the 2d Auxiliary Surgical Group. In two of these injuries, the rib margins had been destroyed by the original wounds, and this approach was used as a matter of expediency. In one of the four other cases, in all of which the incision was elective, the wound broke down and the resulting infection was fatal. Most chest surgeons believed there was nothing to be said in favor of this approach as an elective procedure.3


Management of the Thoracic Component

The thoracic component of the thoracoabdominal wound was treated in much the same manner as if it were the only wound. The principal difference was that, because of the abdominal element and the transthoracic approach, many chest wounds were treated by thoracotomy in field hospitals that otherwise would have been treated in evacuation hospitals and only by debridement and thoracentesis. Once the thoracic cavity was entered, there would have been no point to not removing intrapulmonary foreign bodies and bone fragments, though a prolonged search for small missiles was not warranted. Pulmonary lacerations of more than 4 or 5 cm., as well as any lacerations which demonstrably leaked air, were repaired with interrupted sutures.

3The long vertical abdominal incision introduced by Duval in World War I does not seem to have been employed in World War II. It was used by Wangensteen after World War I, and it was employed by Norwegian surgeons in a forward hospital during the Korean War.-F. B. B.


Management of the Abdominal Component

Standard practices were also followed in the management of the abdominal component of the wound. In brief, wounds of the stomach were sutured. Wounds of the small intestine were treated by suture or by resection and anastomosis, according to the indications. Wounds of the colon were exteriorized or managed by colostomy. Wounds of the spleen were treated by splenectomy. Wounds of the pancreas were treated by suture and drainage or by drainage alone. Wounds of the kidney were preferably treated by drainage; nephrectomy was performed only when the pelvis was definitely involved or when hemorrhage was intractable. The management of all of these wounds is described in detail in the volume on general surgery devoted to abdominal injuries (1).

Wounds of the liver-Wounds of the liver require special mention. Even if only a small fragment had penetrated the liver substance, it was conservative management to explore the wound, since it was not possible to determine from the size of the missile the precise site of the diaphragmatic laceration or the extent of hepatic damage. In some instances, particularly when a rib had been fragmented, a surprising amount of diaphragmatic and hepatic damage was caused by the passage of a small missile.

Experience in World War II soon showed that any wound of the liver that warranted exploration also required drainage. A wound too small to demand drainage was almost never encountered. It is quite true that bile did not drain after operation in some cases, but there were no criteria by which the injuries in which this would happen could be identified at operation. The size of the missile was not the deciding factor. If a small missile cut across a main bile passage, drainage was more profuse than it might be after a more superficial hepatic wound of greater extent. The larger bile radicles were frequently involved in wounds that looked innocent.

If drainage was omitted and infection developed, the resulting draining sinus often took weeks to close. In other cases, the diaphragmatic repair broke down, particularly if the perforation was over the bare area of the liver over which there is no peritoneum. The pooled blood and necrotic material from the damaged liver readily broke through the unfortified diaphragmatic suture line. The resulting biliary pleural fistulas were difficult to treat, and there was more than a casual relation between them and pleural empyemas.

The use of gauze packs was abandoned early in the war as both inefficient and hazardous. They gave rise to subphrenic, subhepatic, or pelvic collections of bile, often with associated abscess formation. Since the diaphragm was perforated in all thoracoabdominal wounds, bile and exuded fluids, whose external discharge was prevented by the pack in the liver, tended to force their way through the sutured diaphragmatic wound, no matter what the technique of suture, with resulting biliary fistulas and bile empyemas. In some instances, the thoracotomy incision broke down, and pleurocutaneous fistulas developed.


In two cases, in which the lung was adherent to the suture line in the diaphragm, a serious problem was created when the bile eroded into a bronchus.

Other disadvantages connected with the use of gauze packs concerned their removal, which was usually painful and was often followed by secondary hemorrhage. In one such case in the experience of the 2d Auxiliary Surgical Group, the granulation tissue that had invaded the meshes of the gauze bled so profusely when the pack was removed that the patient was exsanguinated before surgical intervention could be undertaken. Autopsy showed vascularization of the liver bed as a reaction to the presence of the pack.

Complications associated with the use of gauze packs were so numerous and so troublesome that their use was eventually discontinued, and Penrose drains were substituted for them. These drains, however, were satisfactory only when the drainage incision was of adequate size not only on the surface but through all layers of the drainage tract. It was also essential that they be placed under direct vision when the abdomen was still open.

Adequate drainage implied the proper placing of two or three large, soft rubber drains to the site of the liver damage through an adequate anterolateral subcostal incision. The incision was not adequate unless it was at least 3 inches long. A tract was established extraperitoneally by blunt dissection until the upper surface of the liver was reached; then the peritoneum was broken through. The drains were placed precisely, under direct vision, before the diaphragm was closed. Often both anterior and posterior drainage was desirable. Drains to the liver were never brought to the surface through the thoracotomy incision or the laparotomy incision. This would have been an invitation to wound infection and disruption.

There was a significant reduction in the complications caused by inadequate drainage of wounds of the liver as the principles of drainage became more clearly understood and the proper techniques were put into effect. Constant emphasis upon both principles and practices was, however, necessary throughout the war.

Closure of the Diaphragm

The rent in the diaphragm was repaired as securely as possible to forestall the development of bile empyema on the right side and of diaphragmatic hernias, which were more of a threat on the left than on the right side. When thoracotomy and laparotomy had been done as separate procedures, the diaphragm was preferably repaired from above, through the thoracotomy. Closure from below was much more difficult and much less satisfactory.

The anesthesiologist had much to do with the ease with which the diaphragm was closed. If he initiated a period of controlled respiration, thus securing complete diaphragmatic relaxation and immobility, closure was greatly facilitated. Relaxation was particularly important when the closure had to be effected in a site that would not have been the site of election: in war wounds, the surgeon had no choice.


Several techniques were used to close the diaphragm, a 2-layer closure being generally regarded as most satisfactory. It was accomplished either by imbrication of 2 cm. of diaphragmatic tissue or by simple approximation of the wound edges, reinforced by a second layer of inverting mattress sutures. Neither of these techniques was practical through an abdominal approach.

Crushing of the phrenic nerve was practiced only where there had been actual loss of muscle tissue, so that closure under tension would otherwise be necessary.

Silk or cotton sutures were used, sometimes combined with chromic catgut. Silk was used in 373 of the 636 cases in which 2d Auxiliary Surgical Group records have information on this point. Sutures were always interrupted.

Special techniques were necessary for special wounds. An apparent defect in the lateral costal margin of the diaphragm, which was the result of contraction or avulsion after severance of its attachments, was corrected by suturing the diaphragm to an intercostal muscle one or more spaces higher than its original attachment. Reattachment was important, for the original site of attachment was usually in the area of the chest wall wound, with its accompanying fractures, and this structure was no longer stable enough to permit efficient diaphragmatic action.

When the diaphragmatic crura had been torn from the vertebral column, the first layer of sutures had to be passed through the superior surface of the liver. A sufficient number of these sutures always held to fix the liver against the undersurface of the diaphragm and thus prevent herniation.

Pleural Lavage

Before the chest wall was closed, all blood was suctioned out of the hemithorax, and it was adequately flushed with physiologic salt solution (vol. I). This measure was particularly important if there had been soiling of the cavity by gastric or intestinal contents.


Whether to institute drainage of the pleural space at operation or to rely upon keeping it empty after operation by thoracentesis was a matter of prime importance in securing reexpansion of the lung. The decision depended upon a number of considerations, including the extent of pulmonary damage; the magnitude of the laceration in the diaphragm; the size of the defect in the chest wall; and, most important, whether or not the pleura had been grossly contaminated through the diaphragm from a perforated abdominal viscus.

Unless the defect in the thoracic wall was small, pulmonary and diaphragmatic injuries minimal, and gross contamination of the pleural and peritoneal cavities absent, it was not safe to omit water-seal intercostal drainage and to rely upon postoperative thoracentesis to keep the pleural cavity free of blood and air. When water-seal drainage was employed, the chances of pleural


infection were reduced, and, if it did occur, the process was limited and localized because there was no pleural dead space.

According to the records, closed water-trap intercostal drainage was instituted in 326 of the 903 thoracoabdominal injuries cared for by teams of the 2d Auxiliary Surgical Group. The practice was to use either a large fenestrated catheter or a tube of equal caliber; both were rigid enough not to collapse. The site of drainage was a separate intercostal stab wound, usually in the posterolateral or lateral aspect of the lower chest. A small additional catheter was sometimes placed in the second interspace anteriorly, for the escape of air trapped in the upper pleural cavity.

Closure of the Chest Wall

Closure of the thoracic cage was accomplished by approximation of adjacent ribs or of the intercostal musculature, depending upon whether thoracotomy had been performed through the bed of a rib or through an intercostal space. If there were large defects in the thoracic wall, contiguous musculature was utilized in layers. In closing the posterior aspect of the pleural incision, it was often useful to pull the paraspinalis muscle over the suture line for additional reinforcement. The muscular layer was always closed, but the subcutaneous tissue and skin were usually left open for subsequent closure.


Special surgical conditions in the management of thoracoabdominal wounds depended upon (1) which side they were located, and (2) whether they were penetrating or perforating.

Right-Sided Wounds

Perforating wounds-When the wound on the right side was low in the costal margin or in the region of the tenth, eleventh, or twelfth ribs posteriorly, possible injury of the gallbladder, liver, duodenum, colon, and kidney had to be borne in mind. Exploration by the transdiaphragmatic route was often not satisfactory in injuries in these locations because the liver blocked vision and prevented adequate exploration.

Debridement of the wounds of entrance and exit and investigation of the status of the diaphragm through a traumatic thoracotomy sometimes revealed that only the diaphragm and liver, or sometimes the kidney, had been damaged. If the thoracotomy was as low as the tenth interspace posteriorly, the liver could be drained or the kidney sutured or removed through the diaphragm.

Wounds of the diaphragm no higher than the eighth interspace and no farther posterior than the posterior axillary line could be exposed and sutured from within the abdomen. The condition of the patient determined whether lacerations of the diaphragm in other areas should be repaired at this time.


Failure to repair a laceration of the diaphragm might lead to a biliary pleural fistula, but the possibility of this complication was not of major importance in a patient in such poor condition that he could not tolerate further surgery.

Early surgical intervention was imperative when the right lobe of the liver continued to bleed into the pleural cavity or when the spleen lay shattered in the left pleural cavity. The situation was evaluated on the basis of the location of the wound of entry, the position of the fragment, and the lack of response to shock therapy. It was often necessary to operate without delay in spite of persistent hypotension.

It was also possible to drain a wound of the liver through a traumatic thoracotomy in this location. The most satisfactory plan was to lay Penrose drains over the rent in the liver and exteriorize them subcostally by breaking through the lateral reflection of peritoneum over the right lobe. Whatever the technique used, drainage was essential.

When there was no urgent indication for laparotomy before thoracotomy, the first attention was given to the chest wound. After debridement, closure of the chest wall, insertion of an intercostal tube, and aspiration of all blood and air in the pleural cavity, it was surprising to find how often the patient's condition had improved and how much better he tolerated extensive intra-abdominal surgery. The intercostal tube was drained into a water-trap bottle, and it was thus possible to check on the amount of hepatic bleeding. The drainage of blood into the pleural cavity, which would otherwise have occurred, might have considerably lowered the already impaired vital capacity by the time laparotomy was concluded.

An extrapleural approach was used in an occasional thoracoabdominal injury low on the right side posteriorly, but only if the thoracic injury was not severe and if it was quite certain that the liver was the only abdominal organ injured. The eleventh or twelfth rib was resected, and the reflection of the pleura was dissected up. The pleural defect, being thus exposed, was then repaired, after which the liver was approached either through the perforation in the diaphragm made by the missile or through the bed of the resected rib. The operation was performed exactly as drainage for a subphrenic abscess would be instituted. In selected cases, this was a safe and simple way of draining wounds of the liver.

Penetrating wounds-When a penetrating wound was on the right side, it was necessary to investigate whether the missile had passed through the right lobe of the liver or had entered the peritoneal cavity. As a general rule, the laparotomy approach was best when roentgenologic examination showed the missile to be in the peritoneal cavity. If roentgenologic examination indicated that it had stopped in the liver, the procedure was the same as for right-sided perforating wounds; that is, drainage of the suprahepatic space and repair of the diaphragm. If the fragment was readily accessible, it was removed, but protracted exploration for a buried intrahepatic foreign body was not wise, for it often resulted in considerable hemorrhage. Drainage, as al-


ready pointed out (p. 118), was mandatory to guard against subphrenic space infection and bile empyema.

Left-Sided Wounds

Perforating wounds-Possible injury to the left diaphragm was always in itself a sufficient indication for thoracotomy. It was a simple matter, in the process of handling a sucking wound, to enlarge the wound and determine the status of the diaphragm. Since the chest wall is relatively thin in the lower thoracic cage, most wounds were sucking.

Penetrating wounds-Accurate localization of the missile was important in left-sided penetrating wounds in order to determine the surgical approach. Any damage done to intraperitoneal organs by a missile that entered the left upper quadrant of the abdomen and remained in it could easily be reached and remedied through a transdiaphragmatic approach, with greater technical ease, in fact, than was possible at laparotomy. If, however, the missile had passed into any other quadrant of the abdomen, laparotomy was necessary.

The left transdiaphragmatic approach had many technical advantages. Through it the following procedures (fig. 37) were readily performed: Splenectomy; exploration and repair of the entire stomach from the esophagogastric junction to the duodenum; examination and mobilization of the colon from the hepatic flexure to the first portion of the sigmoid loop; examination of the hepatic flexure itself, by blindly taking down the hepatocolic ligament; repair of the colon, or subcostal or midline exteriorization of either loop, or spur colostomy; examination of the small bowel from the ligament of Treitz to within a few inches of the cecum, with resection or anastomosis as necessary; and visualization and removal of the left kidney.

The abdominal approach was required for repair of wounds of the, lower ileum and cecum and of wounds of the colon from the hepatic flexure to the upper sigmoid. These structures could not be adequately exposed through the diaphragm.

The abdominal approach was also wisest when thoracic damage was so slight that no intrathoracic surgery was necessary. If the diaphragm could be repaired from below in these cases, it was best not to enter another serous cavity.

When the foreign body was in the left upper quadrant of the abdomen and the wound of entrance was not sucking because the missile had traversed the chest wall obliquely, it could be assumed that pleural contamination had not occurred. The decision then had to be made whether to perform laparotomy or to employ a transthoracic approach, with the possible risk of pleural contamination. Many surgeons, on the basis of their personal experience, thought the minimal risk justified because of the greater technical ease of repair of intraperitoneal structures through this approach. The chances of infection were small with an experienced surgeon, who bore in mind the importance of


FIGURE 37.-Technique of transdiaphragmatic operation for left-sided thoracoabdominal wound. A. Exposure of gastric perforations after excision of ninth rib and opening of diaphragm. B. Field after splenectomy and closure of gastric perforations. C. Large laceration in splenic flexure of colon. D. Subcostal exteriorization of splenic flexure as loop (a), or spur colostomy (b). In general, spur colostomy was indicated when more than half of the diameter of the bowel had been transected or when the wound was on the mesenteric side and the blood supply was questionable.

a fully expanded lung, a dry pleural cavity, and proper attention to pleural lavage and pulmonary expansion during closure of the thoracotomy.

The following technique for left-sided thoracoabdominal wounds was reported by Major Shefts and Captain Doud:

After the patient had been anesthetized, he was placed on his right side, with a folded blanket under the lower lateral aspect of the right costal margin. A long strip of adhesive placed over the left hip and attached to the litter anteriorly and posteriorly helped to hold the position. The left leg was flexed at the hip and knee, and a pillow was placed below the knee. With the right leg extended, the great saphenous vein anterior to the medial malleolus or at the middle border of the tibia was available if the administration of fluids during the operation should prove necessary.


The rib chosen for resection varied with the habitus of the patient and could be determined during debridement of the wound. The best approach was usually through the bed of the resected ninth rib. Resection stopped short of the costal cartilage, to prevent the chondritis possible if contamination should have occurred. When the diaphragm was opened, the perforating wound was included in the incision whenever that was possible.

When it was necessary to explore the stomach, access to the posterior wall was obtained by opening the gastrocolic omentum. For exposure of injuries in the region of the esophagogastric junction, especially those on the side of the lesser curvature, it was necessary to cut the left triangular ligament and mobilize the left lobe of the liver.

When splenectomy was indicated, the lateral reflection of peritoneum over the spleen was cut and the spleen delivered into the opening in the diaphragm. The operation was technically easier if, after the spleen had been delivered, the surgeon and his assistant changed sides. The assistant maintained the spleen in position with his left hand, while the surgeon passed his fingers beneath the splenic pedicle. With this maneuver, the risk of injury to the greater curvature of the stomach and the tail of the pancreas was greatly decreased.

To expose the left kidney, the lateral peritoneal reflection of the upper portion of the descending colon was cut, together with the lienocolic ligament, and the colon was retracted mesially. Nephrectomy was readily performed through this approach.


In general, the postoperative care of the casualty with a thoracoabdominal wound consisted of the care that would be given to him if he had a chest wound only (vol. I) plus the management of an abdominal wound only. The routine of the postoperative care of casualties with abdominal wounds is described in detail in the volume of this series dealing with these wounds (1) and need not be repeated here.

As a general rule, the absence of an abdominal incision made for a smoother convalescence because abdominal discomfort and distention were less. Since he was more comfortable, the patient was likely to be more willing to cooperate in the measures necessary for promoting pulmonary expansion and keeping the tracheobronchial tree clear.

Routine of Management

Measures directed to the respiratory system-Before a patient with a thoracoabdominal wound was removed from the operating table, the tracheobronchial tree had been cleared by aspiration or, if necessary, by bronchoscopy. Bronchoscopy was employed in 110 of the 903 thoracoabdominal operations done by surgeons of the 2d Auxiliary Surgical Group. If the patient showed any cyanosis or respiratory difficulty, oxygen was administered at once.


As soon as the patient was conscious, he was given instructions for deep breathing, frequent coughing, and turning from side to side. If he could not or would not cough voluntarily, intercostal nerve block was carried out if it had been omitted at operation.

Tracheal aspiration by catheter was also begun and was repeated as frequently as necessary as long as there were detectable amounts of blood or mucus in the tracheobronchial tree. Often after the first tracheal aspiration, the patient coughed voluntarily and without protest and kept the air passages clear by his own efforts. If he could not or would not cooperate, or if catheter aspiration was unsatisfactory, bronchoscopy was resorted to promptly. It was performed after operation in 12 of the 903 thoracoabdominal wounds cared for by members of the 2d Auxiliary Surgical Group. Atropine gr. 1/100 was given intravenously from 10 to 15 minutes before it was begun, to eliminate the risk of vagovagal reflex. Bronchoscopy could be performed quickly and easily on the ward because most members of thoracic surgical teams-surgeons, assistant surgeons, and anesthesiologists-were all trained to perform it.

If the ward was busy and repeated observation was not practical, it was the best plan to use an oral or nasal airway to maintain patency of the respiratory passages. At least one thoracoabdominal casualty is known to have died of strangulation 3 hours after operation because this precaution was omitted on a busy postoperative ward.

Replacement therapy-Replacement therapy was continued after operation according to the indications. The amount of blood and plasma administered depended upon the hemoglobin, hematocrit, and serum protein determinations. Between 2,000 and 3,000 cc. of 5 percent glucose in physiologic salt solution was given daily. This amount was not usually exceeded because it was still possible to overload the circulation and produce pulmonary edema.

Gastrointestinal decompression-All patients with thoracoabdominal wounds were treated by nasal intubation and gastric siphonage for 3 or 4 days after operation, the removal of the tube depending upon the time at which satisfactory peristalsis was reestablished. The use of the tube beyond this period was undesirable, for it might mask an intestinal obstruction and delay its recognition and correction. An accurate record of the daily intake and output was necessary during the period in which nasogastric suction was in use.

General measures-If morphine was indicated for pain, it was given in small doses, preferably gr. 1/6 and never more than gr. .

Patients were encouraged to get out of bed promptly if other wounds did not contraindicate ambulation. If drainage had been instituted, ambulation usually had to be delayed until the water-seal catheter was removed, but these patients were encouraged to sit on the side of their beds.

The sulfonamides and penicillin were used according to the routine described for thoracic wounds (vol. I).

Vitamins B and C were given routinely in the postoperative period to improve the nutritional status.


The roentgenograms ordinarily taken in chest injuries 4 or 5 days after operation were usually deferred until after the first week in thoracoabdominal wounds. Clinical signs provided a satisfactory guide to management.


The care of water-trap drainage in the postoperative period required constant vigilance on the part of the surgeon and a thorough knowledge of the method plus conscientious attention on the part of the nursing and other staff members of the forward hospital. Serious errors might result when untrained personnel were entrusted with this phase of the patient's care.

The catheter was connected with the water-seal bottle as soon as the patient was returned to the ward. If two tubes were used and penicillin had been left in the pleural cavity, the posterolateral tube was kept clamped for 3 to 4 hours, to give the penicillin time for bacteriostatic action. The patient was watched closely and the tube was unclamped immediately if pneumothorax or hemothorax seemed to be developing.

The water-seal bottle was kept at least 18 inches below the most dependent part of the pleura. The tube was clamped off whenever the bottle had to be emptied. It was the general practice to continue water-trap intercostal tube drainage for 48 to 72 hours unless the tube became sealed off earlier; this sometimes occurred within 24 hours.

After the tube had been removed, and earlier if drainage had not been instituted, aspiration of blood and air from the chest was carried out as indicated. Daily aspirations were sometimes necessary. Often a pocket of air was evident anteriorly, even when drainage had been instituted by lateral or posterolateral intercostal drainage. It was aspirated at once.

The proper removal of drains down to the liver was as important as their initial placement. They were gradually shortened, usually beginning on the fourth or fifth day. Manipulations had to be cautious and gentle. If the drains were withdrawn too far and too fast, fluid collections were likely to become pocketed in the hepatic region.

The amount of drainage through tubes placed down to the liver was generally, though not always, related to the amount of hepatic damage. The period of drainage was unpredictable, but it was seldom shorter than from 10 to 12 days, which meant that the final removal of drainage tubes was usually the responsibility of the surgeon in the base hospital.


Of the patients who survived to reach base hospitals after surgery for thoracoabdominal wounds, about half would have had involvement on the left side and would have undergone, variously, splenectomy; repair of gastric lacerations; exteriorization of a segment of the colon, usually the splenic flexure


or descending colon; nephrectomy in a small number of cases; or various combinations of these procedures.

Of the patients with right-sided injuries, the majority would have had injuries of the liver as the only intraperitoneal injury. The remaining patients with injuries on this side would have had thoracotomy with inspection of the abdomen, repair of the diaphragm, and laparotomy to deal with injuries inaccessible from the right hemithorax.

Except for the closure of traumatic wounds and of elective incisions in which the skin and subcutaneous tissues had been left open, the chief responsibility of surgeons in base hospitals in respect to thoracoabdominal wounds was the management, and sometimes the closure, of colostomies and the management of complications.

Management of Colostomies

The management of colostomies in thoracoabdominal injuries was the responsibility of general surgeons in base hospitals. In many cases, perhaps most cases, the patient had recovered entirely from his chest injuries long before the closure of his colostomy was undertaken. Many patients with thoracoabdominal injuries had to be evacuated to the Zone of Interior because the colostomy could not be closed within the permitted holding period of the theater or because, for other reasons, closure could not be attempted overseas.

There was no standard surgical technique for the management of wounds of the colon in the Mediterranean theater, but all the techniques employed were based on three general principles (1):

1. Exteriorization of the wounded segment of bowel to prevent intraperitoneal leakage at the suture line. The damaged exteriorized segment could be used as the site of colostomy on appropriate indications.

2. Diversion of the fecal stream away from wounds of the distal or lower colon and rectum. Colostomy was always performed for perforation of the rectum and was supplemented by adequate posterior drainage through the fascia propria.

3. Incomplete diversion of the fecal stream, which was a temporary measure, designed for purposes of decompression as well as to bring the bowel to the surface, so that a diversional colostomy could be performed.

As might be expected, when so many different surgeons had operated on so many patients under such widely different circumstances, with only principles to guide them, the colostomies that had to be closed represented every known technique. Closure therefore had to be accomplished by a corresponding variety of techniques. Both intraperitoneal and extraperitoneal methods were found satisfactory. Careful preoperative preparation was a criterion of success.

Techniques of the closure of colostomies and the management of complications associated with them are presented in detail in the volume on general surgery devoted to abdominal injuries (1).



As might be expected in the circumstances of war, the records of complications of wounds were always incomplete and were particularly fragmentary in respect to the less severe and nonfatal types. Many postoperative processes which in civilian practice would be regarded as complications were so frequent in battle wounds that they came to be accepted as routine and were not made a matter of record.

These statements are all applicable to the 903 thoracoabdominal wounds cared for by the 2d Auxiliary Surgical Group. In 529 instances, the records bear no indication that complications occurred, which naturally is incredible. There is only one recorded case, for instance, of postoperative pulmonary edema. The largest numbers of complications were recorded for atelectasis, 21 cases; empyema, 19 cases, 2 of which where associated with bronchopleural fistula; subphrenic abscess, 15 cases; and pneumonia, 12 cases. Ten disruptions of the abdominal wound were recorded, in three instances with evisceration.

Postoperative complications were remarkably few, considering the character of the wounds, but individual surgical experiences often included more than these presumably total figures indicate. Maj. Thomas H. Burford, MC, for instance, in 98 lacerations of the liver treated at base hospitals, collected 14 subphrenic abscesses (fig. 38), 5 bile empyemas, and 6 intrahepatic abscesses (fig. 39).

Right-sided thoracoabdominal injuries gave rise to a higher percentage of complications than left-sided injuries because of the almost universal involvement of the liver on this side. It was estimated that transdiaphragmatic biliary fistulas, with bile empyema, developed in about 3 percent of the lacerations of the liver in right-sided wounds. The incidence of hepatic abscesses was put at about 1 percent.

Almost without exception, patients with left-sided injuries did well if they survived the first week. Complications arising within the abdomen were significantly less than on the right side, though an occasional left-sided subphrenic abscess required drainage.

Bile empyema-Bile empyema resulting from the intrapleural, transdiaphragmatic drainage of bile was always serious. The complication was not usually lethal, but it invariably showed a troublesome predilection toward chronicity, which was attributed to the presence of bile in the pleural cavity.

Attempts at decortication of the lung, with excision of the fistula and tight diaphragmatic repair, failed consistently when they were undertaken as late as 6 or 8 weeks after injury. At this stage, it seemed impossible to achieve complete pulmonary remobilization and total pulmonary reexpansion. On the other hand, early surgical attack, within 3 or 4 weeks after injury, gave gratifying results (p. 130). An important phase of the operation was the provision of adequate extrapleural, extraperitoneal drainage for the bile leakage. This was accomplished by placing the rubber tissue drains exactly at the point of


FIGURE 38.-Thoracoabdominal wound requiring later drainage of anterior subphrenic abscess following laceration of liver.

hepatic injury while the diaphragm was still open and bringing them out through a generous subcostal wound.

Subphrenic abscess-Subphrenic abscesses were almost universally of the anterior type and most frequently developed from an inadequately drained transdiaphragmatic abscess of the liver (fig. 40). They were treated by adequate drainage, which occasionally necessitated resection of a segment of the tenth rib below the pleural reflection. Most patients showed satisfactory progress toward healing of the underlying biliary fistula by the time they were evacuated to the Zone of Interior. In one instance, a left-sided subphrenic abscess ruptured through into the left lower lobe bronchus. Adequate drainage below the diaphragm resulted in prompt and complete cure.

In an occasional case, a bile empyema and a subphrenic abscess coexisted:

Case 1-This patient sustained a penetrating shell-fragment wound of the right chest on 14 April 1945. Exploratory thoracotomy on the same day revealed a penetrating wound of the right half of the diaphragm and a laceration of the liver. Subcostal drainage was instituted, and the diaphragm was closed with interrupted cotton sutures. The immediate postoperative course was uneventful. The subcostal drains were loosened on the fourth postoperative day and removed on the seventh day.


FIGURE 39.-Thoracoabdominal wound requiring later drainage of large intrahepatic abscess.

When the patient was received in the chest center at the 70th General Hospital, Pistoia, Italy, the following day, there was considerable bile-stained fluid in the right chest. The fluid rapidly became purulent and showed persistently large percentages of bile. Later, roentgenologic examination revealed a right subphrenic abscess (fig. 41A).

In preparation for surgery, the patient was given 1,500 cc. of blood in 48 hours, which restored the hematocrit value to 40. He was also given penicillin systemically. Exploration on 28 April, 2 weeks after injury, revealed a large subphrenic abscess and an extensive bile empyema. The one-stage operation performed consisted of drainage of the subphrenic abscess, decortication of the lung, excision of the fistula, and closure of the diaphragm with interrupted silk sutures. Decortication was accomplished without difficulty, and pulmonary reexpansion was prompt and complete. The fistulous opening in the diaphragm, at the site of the previous repair, was excised. With the diaphragm open, the cluster of rubber tissue drains that had previously been placed were carried down to the wound in the liver. The diaphragm was closed in two layers, with interrupted sutures of fine silk. The chest was closed tightly after irrigation of the pleural cavity with physiologic salt solution and the introduction of anterior and posterior intercostal catheters. Penicillin (25,000 units) was instilled into the pleural cavity and left in situ for 6 hours.

The wound healed promptly (fig. 41B), and recovery was uncomplicated and complete (fig. 41C).

Other complications-Bronchobiliary fistulas were infrequent. One patient with this complication died of unexplained, massive retroperitoneal hemorrhage while he was being prepared for surgery.

Complications of left-sided thoracoabdominal wounds were sometimes bizarre. In one instance, a soldier with a previously unrecognized congenital diaphragmatic hernia received a chest wound. The colon was perforated, and what amounted to a traumatic colostomy developed; 3 days after wounding,


FIGURE 40.-Management of subphrenic abscess secondary to laceration of liver. A. Posteroanterior roentgenogram showing abscess on right. B. Detail posteroanterior roentgenogram showing fluid level, air cap, and elevation of right hemidiaphragm.


FIGURE 40.-Continued. C. Lateral roentgenogram showing abscess in characteristic anterior position.

feces were found pouring out through the broken-down thoracotomy wound.

Two patients were observed with gastropleural fistulas. In each instance, the gastric repair had broken down and had caused the breakdown of the diaphragmatic repair. Both fistulas were difficult to close because the skin was so badly excoriated from irritating gastric secretions.

An extremely troublesome complication, which fortunately was uncommon, was the so-called plastic type of peritonitis. Presumably, it was caused by the slow leakage of bile into the general peritoneal cavity. The plastic, nonpurulent exudate that resulted matted the loops of intestine together and gave rise to intermittent bouts of partial intestinal obstruction. Prompt decompression by the Wangensteen or Miller-Abbott technique was the usual treatment. Operation was required only in an occasional case.


The following case histories are illustrative of various aspects of thoracoabdominal wounds:

Case 2-This soldier was operated on in a field hospital, under endotracheal ether anesthesia, 3 hours after sustaining a severe perforating gunshot wound of the left chest, with abdominal involvement. The chest was entered through a left-sided thoracotomy; an 8-inch segment of the ninth rib was resected.

The chest cavity contained the entire stomach, which was tremendously dilated; a portion of the jejunum: the omentum; the transverse colon with the splenic flexure; and the spleen. The stomach, jejunum, colon, and kidney were found to be intact. The badly lacerated spleen was removed. Insertion of a stomach tube reduced the stomach to 40 percent of its original size. The left gastroepiploic artery, which was torn near its termination, was ligated. A tear in the lesser omentum was repaired. After the herniated


FIGURE 41.-Coexistent bit empyema and subphrenic abscess. A. Posteroanterior roentgenogram showing right-sided subphrenic abscess and bile empyema. B. Same, 10 days after one-stage operation consisting of thoracotomy, drainage of subphrenic abscess, decortication bile of lung, and closure of biliary fistula in diaphragm. C. Patient 1 week after operation.


organs were replaced in the peritoneal cavity, the diaphragm was closed in two layers with interrupted silk sutures. The tip of the lower lobe of the left lung, which was badly torn and bleeding, was resected, and the wound sutured with fine catgut on a swaged needle. After an intercostal catheter was inserted, closure was effected with pericostal sutures and suture of the muscle layers. The patient was evacuated to a base hospital on the 12th day in excellent condition.

Comment-This case history typifies the multiple character of left thoracoabdominal injuries because of the close proximity of the organs in the left upper quadrant of the abdomen to each other. The history also demonstrates their ready accessibility through a transthoracic approach. When the diaphragm is opened, they actually rise into the pleural cavity.

Case 3-This soldier sustained a penetrating shell-fragment wound of the tenth left interspace in the posterior axillary line. The wound was explored at a forward hospital, which came under shellfire during the operation. Closure was therefore hurriedly effected after suture of the diaphragm. When the casualty was received at an evacuation hospital 24 hours later, his record contained a note from the forward surgeon that he "thought a laparotomy should be done." The man was in great pain, and his abdomen was considerably distended.

When the abdomen was opened through a left rectus incision, the peritoneal cavity contained fresh blood, but all viscera appeared intact.

Thirty-six hours after operation, the thoracotomy wound became crepitant and edematous, and a foul odor was noted. When the wound was explored, a typical clostridial myositis of the superficial muscles was found. The muscles were already necrotic. Death occurred 12 hours later. Autopsy confirmed the diagnosis of clostridial myositis of the chest wall and also revealed a perforation of the splenic flexure of the colon.

Comment-The injury of the retroperitoneal portion of the splenic flexure was not identified at laparotomy in this case. It would probably have been found, and could have easily been handled, had a transthoracic approach been employed.

Case 4-This soldier was admitted to a field hospital with a rigid abdomen shortly after sustaining a penetrating wound of the right lower chest. Thoracotomy was performed, with resection of the tenth rib, and a laceration of the lower lobe of the right lung was sutured. The diaphragm was perforated and the liver lacerated; the missile lay in the tear. Inspection of the duodenum, hepatic flexure, and kidney revealed no further damage. The wound in the liver was drained subdiaphragmatically before the diaphragm was closed with two layers of interrupted silk sutures. The patient was in good condition when he was evacuated on the sixth postoperative day.

Comment-This case is one of 17 right-sided thoracoabdominal lesions, 12 of which were handled satisfactorily by the transthoracic approach.

Case 5-This soldier was received in a field hospital with a blood pressure of 90/50 mm. Hg and a pulse of 140, after having sustained a gunshot wound of the left chest and abdomen. He had received 500 cc. of plasma before admission. No respiratory difficulty was apparent, and breath sounds were present, though somewhat diminished on the left. The abdomen was spastic and moderately tender. He complained of suprapubic pain.

The wound of entrance, which was 1.5 cm. in diameter, was in the seventh interspace anteriorly, 2 inches from the midline. The ragged wound of exit, 7.5 cm. in diameter, was in the ninth interspace, just medial to the posterior axillary line. Aspiration of the left chest produced about 30 cc. of blood and a small amount of air.

After a plasma transfusion of 500 cc. had been given, a blood transfusion was started and ether-oxygen anesthesia induced. A subcostal laparotomy incision was made. Exploration revealed approximately 800 cc. of blood, mixed with gastric contents, in the peritoneal cavity. The stomach was greatly distended. The lesser peritoneal cavity was not involved. The spleen was lacerated, and a perforating wound was found in the cardia.


The wounds of entrance and exit in the cardia were closed in two layers with chromic catgut by the inverting technique. When the spleen was delivered, a 6-cm. sucking laceration of the posterior portion of the diaphragm became evident. It was closed with difficulty after the spleen had been removed. Another nonsucking laceration of the diaphragm was easily repaired. The abdomen was closed in layers.

The patient's condition was not satisfactory at any time during the 90-minute procedure, and he died while the skin sutures were being inserted. At no time did the blood pressure rise above 90/60 mm. Hg, in spite of the continuous administration of blood, and there was a gradual decrease in pressure during the last 30 minutes of the operation.

Comment-The first error made in this case was failure to have an endotracheal tube in place before the abdomen was opened. The large sucking perforation of the diaphragm was not apparent until the lacerated spleen was delivered and its tamponade effect on the diaphragm was lost. Except for the application of an ineffective gauze pack, the diaphragmatic laceration had to be ignored until splenectomy was completed. Most of the operating time was spent in accomplishing a difficult diaphragmatic closure. The sudden collapse of the left lung was more than an already shocked patient could tolerate. Had endotracheal anesthesia been used, with facilities for maintaining positive intrapulmonic pressure during operation, pulmonary collapse would not have occurred. The major error of management was not to employ the transthoracic approach, through which the entire procedure could have been carried out with a great reduction in risk.


The World War II experience of the 2d Auxiliary Surgical Group in the Mediterranean theater included 903 thoracoabdominal wounds, 247 of which were fatal. These wounds were encountered chiefly during 1944 and 1945, but 68 wounds, 25 of which were fatal, were observed in 1943. During the same 1943-45 period, the teams of this group encountered 1,364 thoracic wounds. Thoracoabdominal wounds thus accounted for almost 40 percent of all wounds of the chest.

Certain data concerning these 903 thoracoabdominal wounds have been cited in the preceding pages. Additional data, and certain conclusions which can be drawn from them, follow.

General Data

The great majority of these casualties were U.S. Army infantrymen, with Wehrmacht prisoners of war accounting for the next largest group. The figures also cover some British and French soldiers and a few civilian casualties.

The age factor was of little significance, as the great majority of the patients were young, healthy males. In the 762 cases in which the age was recorded, there were only 4 patients over 40 years of age. There were 190 casualties 20 years of age and under; 297 in the 21- to 25-year group; 176 in the 26- to 30-year group; and 95 in the 31- to 40-year group.

Artillery and mortar fire was responsible for 590 of the 903 wounds, and rifle, machinegun, and pistol fire for 245. The remaining wounds were caused by bombs, mines, and grenade fragments.


In 837 cases, the missiles entered the abdomen through the thorax. In the other 66 cases, they entered the abdomen and then penetrated the diaphragm. In 418 cases, the wounds were on the right side and in 419 on the left side. In 20 cases in which both diaphragms were perforated, the wound of entry was on the right in 7 and on the left in 13. No information is available on this point in the remaining 46 cases.


There was no record of the timelag in 49 cases, 11 of which were fatal. In the 854 cases in which information on this point is available, it varied from less than an hour to more than 30 hours (table 8).

At first glance, the high mortality rate in the group in which the timelag was shortest might seem to invalidate the concept that the earlier a casualty with a thoracoabdominal wound could be operated on, the better were his chances. It does not. The high rate reflects the fact that the nearer the frontline a field hospital is set up, the more casualties will be brought to it who will die there rather than on the battlefield. This was particularly true on the Anzio beachhead, when both field and evacuation hospitals were very far forward because the sector occupied was too small for them to be placed farther back. On the beachhead, one patient died of an injury of the portal vein, and five died of injuries of the vena cava, in hospitals. If the evacuation distance had been longer, they would not have reached them alive.

There were 8 deaths in the 34 casualties who came to operation 30 hours or more after wounding. Two had wounds of the liver, three wounds of the spleen, and one a wound of the kidney. In the two remaining cases, the missile penetrated the peritoneal cavity without damage to any viscera.

Surgical Approach

The transthoracic approach was used in 448 of these 903 thoracoabdominal injuries (table 9). In 86 patients in this group, surgery was limited to thoracotomy, exploration, removal of foreign bodies, or simple retroperitoneal

TABLE 8.- Case fatality rate according to timelag in 854 thoracoabdominal injuries1





To 6 hours




To 12 hours




To 18 hours




To 24 hours




To 30 hours




Over 30 hours








1In 49 of the 903 cases in this series, there was no record of the timelag. There were 11 fatalities in this group.


TABLE 9.-Case fatality rate according to surgical procedure in 903 thoracoabdominal wounds





Transdiaphragmatic laparotomy








Thoracotomy and laparotomy




Laparotomy and thoracotomy




Transdiaphragmatic laparotomy and laparotomy








No surgery




Death during surgery








1 In this technique, the thoracic incision was extended onto the abdomen over the chondral arch.

drainage. In the remaining 362 cases, the entire abdominal procedure, which involved surgery on 31 different combinations of organs, was performed through the chest incision.

In 202 other cases, the entire procedure was abdominal.

A comparison of the case fatality rates in the injuries handled by the transthoracic approach and those handled by the abdominal approach, 20.3 percent versus 38.1 percent, might suggest the far greater superiority of the transthoracic approach. These data should not be so construed. The results in the individual case depended upon the magnitude of the abdominal and thoracic problems (table 7). Generally speaking, however, the results were better if the thoracotomy approach was used for perforating wounds in the upper zone of the left diaphragm or penetrating wounds in which roentgenograms showed that the missile lay within the area of the dome of the diaphragm.

Factors of Mortality

General considerations-The gross case fatality rate for the 903 thoracoabdominal injuries cared for by surgeons of the 2d Auxiliary Surgical Group was 27.4 percent (247 deaths). These figures are for first priority hospitals, in which the postoperative stay, though it varied from a few days to as long as 18 days, was usually from 7 to 10 days.

Conditions in theaters of operations made it extremely difficult to determine the end results of forward surgery. Reports from evacuation and base hospitals were all to the effect that patients operated on in field hospitals did well after they reached hospitals farther to the rear. It is probable, therefore, that a more complete followup would not have materially altered the case fatality rates in forward hospitals.

There was an interesting and probably significant decrease in the case fatality rates of thoracoabdominal injuries as time passed:

In 1943, there were 25 deaths in 68 cases, 36.8 percent.


Between January and June 1944, there were 84 deaths in 243 cases, 34.6 percent.

Between July and December 1944, there were 98 deaths in 392 cases, 25.0 percent.

Between January and the first week in May 1945, when the war ended, there were 40 deaths in 200 cases, 20.0 percent.

The decreasing case fatality rate in the latter half of 1944 and 1945 coincided with the introduction of penicillin. Antibiotic therapy undoubtedly played a part in the improvement. Another explanation of the improvement is applicable to the first half of 1944, the establishment of a theater blood bank and the increasingly liberal use of whole blood.

Surgeons in the theater, however, believed that the improvement in the case fatality rate could reasonably be assumed to be a reflection of their greater knowledge and experience in dealing with these wounds; a greater appreciation of the implications of the thoracic component and the complications which could arise from it; a greater facility on the part of individual surgeons and anesthesiologists (the so-called learning curve); and a keener appreciation of, and greater attention to, details of resuscitation and of preoperative and postoperative care.

Effect of thoracic and abdominal components of wound-When the mediastinum was injured, the case fatality rate was 54 percent. Otherwise, it was impossible to discover, from a study of these cases, exactly what effect the chest component of the wound had, beyond the far-reaching influence of cardiorespiratory disturbances that were not promptly corrected.

With these exceptions, there was no doubt that the most lethal part of the wound was the abdominal component. The precise organs injured and the number of injuries below the diaphragm were reflected in the mortality rates (table 7). Wounds of hollow viscera carried a much higher rate than wounds of solid viscera, though an occasional fatality was caused by an overlooked wound of a solid viscus.

The number of organs injured also played an important role. Wounds of five or more viscera were universally fatal. These subjects are discussed in more detail in the volume in this series devoted to abdominal injuries (1).

Anatomic considerations.-The side on which the injury was located played a part in the mortality rate, for anatomic reasons. The figures are as follows:

Perforations of the right diaphragm, 435, with 103 deaths (23.6 percent).
Perforations of the left diaphragm, 448, with 136 deaths (30.8 percent).
Bilateral perforations, 20, with 8 deaths (40 percent).

Anatomic reasons are also connected with the case fatality rates for herniation of the abdominal viscera into the chest cavity. In the 5 instances of evisceration into the right pleural cavity there were 2 deaths, against 13 deaths in 35 herniations through the left diaphragm.


Time of death-The day of death was recorded in 234 of the 247 fatalities:

1. There were 154 fatalities in the first 48 hours, including 49 deaths on the operating table (40 from shock, 4 from hemorrhage, 3 from atelectasis, and 1 each from cardiac tamponade and vagovagal reflex during bronchoscopy). All but 13 of the 130 deaths from shock occurred within this period. The case fatality rate could be directly related to the degree of shock in which the casualties were received:

No signs of shock, 145 patients, 8 deaths (5.5 percent).
Mild shock, 75 patients, 10 deaths (13.0 percent).
Moderate shock, 174 patients, 31 deaths (17.5 percent).
Severe shock, 275 patients, 164 deaths (59.6 percent).
There is no record of the degree of shock in the remaining 234 cases, in which there were 34 deaths (14.5 percent).

In addition to shock, chiefly due to blood loss, deaths early in the postoperative period were caused by uncorrected (occasionally irreversible) disturbances of cardiorespiratory physiology, overwhelming contamination and infection of the peritoneal and pleural cavities and retroperitoneal space, and massive tissue destruction.

2. There were 28 deaths between the third and fifth days. Of these, 13 were due to renal failure, as evidenced by oliguria and anuria (vol. I); peritonitis and pneumonia were the next most frequent causes.

The 22 deaths from peritonitis occurred between the day of operation and the 10th postoperative day. The fact that 17 had occurred by the end of the third day is evidence of the catastrophic effect of the so-called contamination type of peritoneal infection, with sudden, massive soiling of the cavity.

3. There were 10 deaths from pulmonary embolism, all between the day of operation and the 11th postoperative day; 6 occurred within the first 72 hours.

Other occasional causes of death included missed perforations of hollow viscera, mismatched blood, blast injury, mediastinitis, empyema, fat embolism, pressure pneumothorax with bronchopleural fistula, liver death, clostridial myositis, disruption of the diaphragm, morphinism, and meningitis.

Associated wounds-Wounds in other parts of the body accounted for some of the 247 deaths in this series or played an important part in the fatality. Of the 385 associated wounds, 161 were classified as mild, and the 34 deaths in this group (21.1 percent) can probably be attributed entirely to the thoracoabdominal wounds. The same reasoning perhaps holds for the majority of the 129 associated injuries of moderate severity, in which there were 33 deaths (25.6 percent). In the 95 associated injuries classified as severe, there were 36 deaths (37.9 percent). There can be no doubt that in many instances in this group the associated injuries played an important role, if not the major role, in the fatality. Among these 95 severe associated wounds there were:

26 cord injuries, with 16 deaths.
7 brain injuries, with 1 death.


15 fractures of the femur, with 4 deaths.
9 traumatic amputations, with 3 deaths.
8 other abdominal (not thoracoabdominal) wounds, with 4 deaths.

The influence of severe associated wounds is evident in a casualty cared for by Maj. (later Lt. Col.) Reeve H. Betts, MC. This was the only casualty with a thoracoabdominal wound to die before operation in the 29 such injuries handled by his team. This soldier had a right traumatic thoracotomy, a left thoracoabdominal wound, a traumatic amputation of the right arm, and a severe wound of the right thigh. In spite of vigorous attempts at resuscitation, including the administration of 500 cc. of plasma and 3,500 cc. of blood, his blood pressure was never obtainable, and he died 5 hours after wounding.

It might be added that associated wounds, in addition to the part they played in the case fatality rate of thoracoabdominal wounds, also played a major role in their morbidity.4


1. Medical Department, United States Army. Surgery in World War II. General Surgery. Volume II. Washington: U.S. Government Printing Office, 1955.

2. Snyder, Howard E.: The Consultant in Surgery, Fifth U.S. Army. In Medical Department, United States Army. Surgery in World War II. Activities of Surgical Consultants. Volume I. Washington: U.S. Government Printing Office, 1962, pp. 333-464.

4 The reader is referred to chapter XI (p. 441) for long-term followup studies on casualties with thoracoabdominal wounds.