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

ACCESS TO CARE External Link, Opens in New Window

















Special Reports and Statistical Data

B. Noland Carter, M.D., and Michael E. DeBakey, M.D.

Throughout this volume, statistical and other material has been cited from the clinical experiences in World War II of various organizations and individuals. This appendix summarizes the clinical-statistical data on thoracic surgery of the activities of:

1. The 2d Auxiliary Surgical Group in the Mediterranean (formerly North African) Theater of Operations, U.S. Army, from the final report made by Maj. (later Lt. Col.) Reeve H. Betts, MC; Maj. (later Lt. Col.) Paul C. Samson, MC; Maj. Lyman A. Brewer III, MC; Maj. (later Lt. Col.) Lawrence M. Shefts, MC; and Maj. Thomas H. Burford, MC.

2. The 5th Auxiliary Surgical Group in the European Theater of Operations, U.S. Army, from the final reports on 1,068 war wounds of the thorax and abdomen, made by Maj. William H. Falor, MC; Maj. Charles B. Burbank, MC; and Maj. (later Lt. Col.) Elmer D. Gay, MC.

3. The Kennedy General Hospital thoracic surgery center, Memphis, Tenn. The material used is an analysis of the first 500 thoracic casualties received in the center after it became fully operational. The report was prepared by Maj. Earle B. Kay, MC, and Lt. Col. Richard H. Meade, MC.

The original detailed reports are on file in The Historical Unit, U.S. Army Medical Service.


Basic Data

Between the Tunisian landings in November 1942 and the end of the war in Italy in May 1945, the 27 general surgical and 5 thoracic surgical teams of the 2d Auxiliary Surgical Group1 cared for approximately 22,000 battle casualties in North Africa, Sicily, Italy, southern France, and Germany. The 3,154 abdominal injuries included in this material are analyzed in another volume of this series.2

During the specified period, surgeons of the 2d Auxiliary Surgical Group cared for 1,364 casualties with thoracic wounds, 135 of whom (9.89 percent) died, and 903 casualties with thoracoabdominal wounds, 247 of whom (27.35 percent) died. This is a total of 2,267 casualties with thoracic and thoracoabdominal wounds and 382 deaths (16.85 percent).

Among the casualties with only chest wounds were 115 civilians and prisoners of war, of whom 19 (16.52 percent) died. This leaves 116 deaths in 1,249 U.S. Army personnel with thoracic wounds (9.29 percent). The following discussion includes only these 1,249 casualties unless otherwise specified.

The figures include all penetrating and perforating wounds encountered by the group surgeons during this period. They do not include injuries limited to the thoracic cage,

1The extensive experience of the teams of the 2d Auxiliary Surgical Group with thoracoabdominal injuries is summarized under that heading (p. 136).
2Medical Department, United States Army, Surgery in World War II. General Surgery. Volume II. Washington: U.S. Government Printing Office, 1955, pp. 79-333.


without pleural penetration, even though pleural damage sometimes occurred under these circumstances.

In the early days of the Tunisian campaign, a small number of these casualties were operated on in clearing stations. A few were also operated on in beach installations, during the first few days of amphibious landings. With these exceptions, all the casualties received their initial wound surgery in field and evacuation hospitals.

All the deaths occurred in these forward hospitals, for which the figures can be assumed to be complete. On the other hand, casualties were held in forward hospitals only for limited periods, seldom more than 14 days. It must therefore be assumed that a certain number of deaths occurred later, in fixed hospitals and thoracic surgery centers in the base. There is no information on this point, for the attempts at followup that were occasionally made were not successful.

Deficiencies of records-Many errors and imponderables enter into any records prepared by medical officers under field conditions. There are errors and omissions in these. Yet they are surprisingly complete when one considers the circumstances in which the material was compiled. It was necessary for each team to prepare from four to seven official records on each patient, all in longhand, all without secretarial or stenographic help, and all containing much duplicated material. All post mortem examinations were made by members of the teams, and all protocols were written by them in longhand. During rush periods, when surgeons were working long hours under extremely adverse conditions, their zeal in the preparation of additional records for some future study was naturally limited. The most amazing thing about the records is that they were kept at all.

These statistics and conclusions are based on the overall experience of the surgeons of the 2d Auxiliary Surgical Group. They represent not only individual experiences but the aggregate experiences of all the surgeons. The number of injuries is large enough to warrant conclusions. The casualties were treated over a sufficiently long time, in sufficiently varied climates and terrains, under sufficiently varied conditions, to compensate for some of the variables and to cancel out some of the inevitable errors. The chief justification for the presentation of this clinical-statistical material in detail is that, so far as is known, this is the most extensive study of combat-incurred thoracic and thoracoabdominal wounds that has ever been made.

Chronologic division of cases-There are a number of reasons for analyzing certain of these data according to whether the casualties were treated before or after 1 May 1944. The date, which is arbitrarily selected, represents a turning point in clinical policies, to which the improved results (a drop in the case fatality rate from 11.27 percent to 8.3 percent) bear witness. Among the factors responsible for this improvement were:

1. Better triage, based on a clarified definition of transportability (vol. I).

2. More rapid evacuation of casualties from the battlefield to the division clearing station.

3. A more correct appreciation of the physiologic components of resuscitation, including the importance of wet lung (p. 207).

4. A clarification of the indications for thoracotomy in forward hospitals and of the limitations of this operation.

5. Improvements in postoperative care.

6. The provision of liberal amounts of blood by a theater blood bank, so that every casualty could be given all he needed.

7. Improvements in anesthesia, as a result of both more adequate equipment and the increased experience of anesthesiologists.

8. The universal availability of penicillin, which after 1 May 1944 could be used in adequate amounts whenever it was indicated.

9. The increased experience of both thoracic surgeons and general surgeons who were called upon to care for thoracic casualties, which is discussed elsewhere as the so-called learning curve (p. 139).


It was the opinion of many surgeons of the 2d Auxiliary Surgical Group that casualties cared for after 1 May 1944 were, generally speaking, more severely wounded than those treated before that date. This is not a matter which lends itself to statistical proof. If the opinion were correct, and if other factors had remained equal, a higher case fatality rate might have been anticipated after that date. Instead, as just pointed out, the rate fell, probably for the reasons just specified.

Wounding Agents and Location of Injuries

In the 1,203 injuries in which this information is available, the wounding agent was a shell fragment in 830 cases, a gunshot wound in 371 cases, and a stab wound in 2 cases.

In 1,240 stated injuries, the right side was involved 602 times (48.55 percent) and the left side 603 times (48.63 percent); both sides were involved 35 times (2.82 percent).

In 1,238 injuries, 719 of the wounds were penetrating (58.08 percent), 478 perforating (38.61 percent), and 41 lacerating (3.31 percent).

Shock and Resuscitation

Shock-Of the 400 thoracic casualties treated before 1 May 1944 (hereafter referred to as Group I) for whom this information was recorded, 149 were not in shock, 67 were in mild shock, 83 were in moderate shock, and 44 were in severe shock. The corresponding data for the 849 casualties treated after 1 May 1944 (hereafter referred to as Group II) indicate that 314 were not in shock, 162 were in mild shock, 190 were in moderate shock and 110 were in severe shock.

Blood and plasma-During the early fighting in North Africa, plasma was used liberally and blood infrequently. Blood was used in increasing amounts and on broadening indications in the fighting in Sicily and in the first months of fighting in Italy, but the supply was limited, and it was often difficult to obtain until a theater blood bank was established in Naples in February 1944. That circumstance is part of the explanation for the more limited use of blood and the more liberal use of plasma in Group I injuries in this series. On the other hand, there must also be taken into consideration that in the early months of the war, it was believed that plasma was an acceptable substitute for whole blood and that in some quarters, at least, this belief died slowly. The amount of blood and plasma administered to these two groups of casualties therefore depended not only upon their special necessities but upon the availability of whole blood and the original impression of the value of plasma.

It is not thought that the statistics that follow are complete, though they are based on all the specific information recorded on the emergency medical tag and on the records of the hospital in which initial wound surgery was performed. Entries on some additional charts suggested that replacement therapy might have been employed, but the data were not specific and were discarded.

The composite recorded figures for replacement therapy are as follows:

In Group I, 196 or 400 patients (49.00 percent) received plasma in the total amount of 138,200 cc., the average amount per patient being 705 cc. The largest amount given was 2,500 cc.

In Group II, 569 of the 849 casualties (67.02 percent) received a total amount of 351,200 cc. of plasma, the average amount per patient being 617 cc. The largest amount given was 2,750 cc. The drop in Group II of almost 100 cc. in the average per case probably reflected the increased availability of blood at this time and the restriction of plasma to its proper uses.

In Group I, 172 of the 400 patients (43.00 percent) received a total of 179,900 cc. of blood, the average per patient being 1,046 cc. The largest amount given was 3,600 cc.


In Group II, 633 of the 849 casualties (74.56 percent) received a total amount of 753,850 cc., the average amount per patient being 1,189 cc. The largest amount given to a single patient was 7,500 cc.

Autotransfusion was used in 29 patients in Group I, in the total amount of 23,400. The average per patient was 807 cc., and the largest amount thus given was 2,700 cc.

In Group II, 42 patients were treated by autotransfusion, in the total amount of 28,950 cc. The average per patient was 689 cc., and the largest amount given was 2,000 cc. The figures for Group II probably reflect the increased availability of blood; autotransfusion is not a desirable technique if blood can be secured elsewhere.

Other resuscitative measures-In Group I, 8 of the 400 patients were treated by intercostal nerve block, 53 by thoracentesis (in 1 case via intercostal catheter), and 4 by tracheal aspiration per catheter. Preoperative bronchoscopy to clear the airway was not used in any instance in Group I.

In Group II, 69 of the 849 casualties were treated by intercostal nerve block, 218 by thoracentesis (in 9 cases by intercostal catheter), 25 by tracheal aspiration per catheter, and 8 by bronchoscopy. One must be impressed by the more active measures used in Group II to facilitate deep breathing and coughing and to keep the airway clear. In spite of the time consumed by these additional measures, the average timelag fell from 15.7 hours in Group I to 14.0 hours in Group II.

Measures to control pressure pneumothorax were seldom necessary. This complication was encountered only once in Group I and only 10 times in Group II.

Surgical Procedures

In Group I, 234 of the 400 patients (58.50 percent) were treated only by debridement. In Group II, 534 of the 849 patients (62.90 percent) were thus treated.

In Group I, 67 thoracotomies were performed by general surgical teams in the 243 cases which they handled (27.57 percent). There were 11 deaths (16.42 percent). In this same group, over the same period, 72 thoracotomies were performed in 157 cases (45.86 percent) by thoracic surgical teams, with 10 deaths (13.89 percent).

In Group II, 224 thoracotomies were performed in 623 cases by general surgical teams (35.95 percent), with 27 deaths (12.05 percent). In this same group, 72 thoracotomies were performed in 226 cases by thoracic surgical teams (31.86 percent), with 6 deaths (8.33 percent). The casualties in this group were all cared for after strict limitations had been put upon the use of thoracotomy. This fact is reflected in the much smaller proportion of cases in which the operation was used, as well as in the reduced case fatality rates.

Thoracotomy was performed through the wound in 368 cases, in 156 of which it was necessary to enlarge the traumatic thoracotomy, and by a separate elective incision in 67 cases.

The indications for the operation, according to the 144 charts on which this information was available, were as follows:

Combined thoracic and abdominal wounds (as contrasted to thoracoabdominal wounds) in 63 cases, with 13 deaths.

Bleeding in 36 cases, with 6 deaths.

Injuries to the mediastinum, heart or both in 20 cases, with 2 deaths. 

Retained foreign bodies in 11 cases.

Retained bone fragments in 3 cases, with 1 death. 

Pulmonary lacerations in 2 cases, with 1 death. 

Bronchial injuries in 3 cases, with 1 death. 

Injuries of the esophagus in 3 cases, with 3 deaths.

Possible injury of a hilar vessel, pressure pneumothorax, and failure of the lung to expand in 1 case each. The casualty with the vascular injury did not survive.


Analysis of Fatalities

The breakdown of the 116 deaths in 1,249 wounds of the chest in U.S. Army personnel (9.29 percent ) was as follows:

53 deaths in 768 debridements (6.90 percent).

19 deaths in 234 debridements in Group I (8.12 percent). 
34 deaths in 534 debridements in Group II (6.37 percent). 

54 deaths in 435 thoracotomies (12.41 percent).

21 deaths in 139 thoracotomies in Group I (15.11 percent). 
33 deaths in 296 thoracotomies in Group II (11.15 percent). 

9 deaths before, or in the course of, operation.

There were 19 deaths in the 115 wounds in civilians and prisoners of war not included in this analysis (16.52 percent).

Among the 1,249 casualties with chest injuries were 200 who also had associated wounds of major importance, including 63 wounds of the abdomen, 42 severe compound fractures, 39 spinal cord injuries, 15 severe soft tissue wounds, 7 traumatic amputations, and 4 contralateral thoracoabdominal wounds.

There were 46 deaths in these 200 cases (23.0 percent), as compared to 70 deaths in the 1,049 U.S. Army casualties (6.67 percent) whose wounds were limited to the chest or who had associated wounds not of major importance. In all combat injuries, the risk was always greater when the wounds were multiple, and it increased in proportion to the severity of the principal and associated injuries. In many of the cases just listed, the chest wound was obviously of lesser importance than the associated injury.

Causes of Death

In 16 of the 116 deaths in U.S. Army personnel, the cause was not recorded and it was impossible to gather from the scanty notes to what they should be attributed. The 41 separate causes listed for the remaining 100 deaths can be conveniently grouped as follows:

Pulmonary and respiratory causes-Of the 100 deaths, 28 were due to pulmonary or respiratory conditions. This was the largest single group of deaths, and, from the clinical standpoint, the group most amenable to treatment. In the opinion of the surgeons who cared for these casualties, some of these deaths, at least, must be regarded as preventable. 

Seven deaths from pulmonary edema occurred between the first and fourth postoperative days. In every instance, the pulmonary damage was severe. It is not known whether positive pressure oxygen therapy was employed in any of these patients, but in several cases, it was thought that injudicious replacement therapy played a part in the fatality.

In 15 cases, the cause of death was listed as pneumonia. In some instances, exposure on the battlefield, preexisting upper respiratory infection, and the virulence of the organism in relation to the resistance of the host explained both the complication and the fatal outcome.

Four patients died of atelectasis. One died of tracheal obstruction caused by an excessive production of mucus, and another died of wet lung, on the fifth postoperative day. Attempts to improve drainage of the tracheobronchial tree were always vigorous, once the importance of the wet lung of trauma had been realized (p. 207), but occasionally they were not successful. Sometimes there was a recurrence of the tracheobronchial obstruction after it had been relieved, and the stage was then set for the development of atelectasis and pneumonia if prompt, zealous efforts were not made to overcome it.

Shock.-Ten patients died of shock, eight on the day of operation, one on the following day, and one on the second postoperative day.


Nine patients died of renal failure, which in most instances seemed related to their previous state of shock. All these deaths occurred in Group II; earlier in the war, lower nephron nephrosis was not recognized and, as a matter of fact, much about it remained to be clarified when the war ended.

Hemorrhage-Nine patients died of hemorrhage, as follows:

3 from lacerations of the aorta.
2 from massive intrapleural hemorrhage.
2 from massive intrapulmonary hemorrhage. 
1 from mediastinal hemorrhage.
1 from laceration of the vena cava.

The small number of deaths from intrapleural and intrapulmonary hemorrhage (only five when the lacerations of the great vessels are excluded) is striking in view of the emphasis put upon intrathoracic bleeding as a possible cause of death in the prewar discussion of chest injuries.

Cardiac deaths-There were five deaths from cardiac injuries, three of which had been overlooked, and another death from cardiac tamponade. One of these injuries was complicated by atelectasis. Two other deaths were caused by cardiac fibrillation.

Miscellaneous causes of death-Other causes of death, which cannot be conveniently grouped, were as follows:

6 deaths from blast injuries, which affected the pulmonary tissues in 4 cases, the mediastinal tissues in 1 case, and the cerebral tissues in 1 case.

4 deaths from clostridial myositis, in 1 instance complicated by pneumonia. 

4 deaths from pulmonary embolism.

3 deaths from mediastinitis, in all of which injuries to the esophagus or the esophagus and trachea had been overlooked.

3 deaths from cerebral anoxia.

2 deaths from aspiration pneumonia, both caused by aspiration of vomitus during anesthesia.

1 death from each of the following causes: cerebral malaria, cerebral embolism, cerebral abscess, phosphorous burns, splenomegaly and jaundice, severe lacerations of the lung, the vagovagal reflex, peritonitis, thrombosis of the pulmonary artery, empyema, right heart failure, massive emphysema, pressure pneumothorax, and multiplicity of wounds. 

Post mortem observations.-Post  mortem examination was carried out in 78 of the 116 fatalities, the most impressive findings being 12 overlooked injuries, including 3 perforations of the aorta; 3 cardiac injuries, 1 associated with massive pulmonary collapse; 3 injuries of the esophagus, 1 associated with an injury of the trachea; 1 injury of the trachea; and 1 injury of the vena cava. Thoracic surgical teams cared for only 3 of these 12 cases.

Chronology of deaths-Of the 116 deaths, 9 (7.76 percent) occurred before surgery could be undertaken or on the operating table. Of the remainder, 26 occurred the day of operation, 19 on the first postoperative day, and 9 on the second postoperative day. In other words, something over half (54.3 percent) of the 116 fatalities had occurred by the end of the second postoperative day. All but 10 of the 116 fatalities (91.4 percent) had occurred by the seventh postoperative day.


In the 1,068 combat-incurred wounds of the chest and abdomen treated by surgical teams of the 5th Auxiliary Surgical Group in the European theater in 1944 and 1945 were 374 penetrating and perforating wounds of the chest, with 31 deaths (8.29 percent) and 165 thoracoabdominal wounds, with 39 deaths (23.70 percent). A few of these wounds occurred in civilian casualties, but the analysis does not distinguish them from the injuries in military personnel.


Thoracic Wounds

Basic data-Of the 374 patients with perforating or penetrating wounds of the chest, 283 were treated in field hospitals and 91 in evacuation hospitals. Field hospitals received twice as many casualties within the first 12 hours after wounding as in the second 12-hour period. Evacuation hospitals received about the same number in each 12-hour period. The timelag for field hospitals averaged 7 hours and for evacuation hospitals, 10 hours. 

Of these wounds of the chest, 180 were on the left and 189 on the right; 5 were bilateral. Gunshot caused 193 wounds and shell fragments 181. Penetrating wounds caused by shell fragments were twice as frequent as perforating wounds, and sucking wounds were caused by shell fragments twice as frequently as by gunshot. Sucking wounds were also twice as common in penetrating as in perforating wounds.

Policies of management-In the European theater, as in the Mediterranean theater, the management of chest wounds progressed through a definite cycle. In the beginning, before surgeons had had much experience with combat-incurred wounds, the general tendency was to be somewhat radical, and there were few patients with chest wounds who did not undergo thoracotomy. When casualties became heavy, however, and there was a large backlog of patients with abdominal and thoracoabdominal wounds, experience showed, as it had in the Mediterranean theater, that casualties with thoracic injuries could be treated satisfactorily by delayed surgery and by less radical procedures.

The following indications for thoracotomy then became regulation:

1. Hemothorax due to active bleeding, with rapid refilling of the pleural cavity after aspiration. The source of bleeding was usually an intercostal artery or the internal mammary artery; the lung itself; or, less often, the heart or some other mediastinal structure. Patients in this group were usually in severe shock on admission and did not respond well to resuscitation.

2. Massive clotted hemothorax that did not lend itself to aspiration. These patients often presented dyspnea, cyanosis, and mediastinal shift. Evacuation of clotted blood and control of bleeding were readily effected through an open thoracotomy, and this method of management became the procedure of choice.

3. Retained foreign bodies 2 cm. or larger in the pulmonary parenchyma or the pleural cavity. The principal indication for removal of these objects was risk of an infected hemothorax or lung abscess. Possible future damage to the lung or to a blood vessel was also considered an acceptable indication.

4. Shattered rib fragments in the lung or pleural cavity or large fragments that had not perforated the pleura. Many surgeons came to believe that bone fragments in the lung were of much more serious consequence than metallic foreign bodies. They were apt to be long and spiky, with irregular edges, and they tended to penetrate the tissues and to forge ahead in them. After a time, bone fragments often became necrotic, and an abscess or an infected hemothorax was a possible consequence. Laceration of a blood vessel was also far more likely from contact with bone fragments than from contact with metallic foreign bodies. The preferable technique of removal was thoracotomy or extrapleural rib resection.

Sucking wounds. If intrathoracic damage of any consequence was suspected, thoracotomy was performed. Otherwise, simple debridement and closure was usually all that was necessary.

6. Tension pneumothorax due to an air leak, as in bronchopleural fistula.

After these indications had been set up, the treatment of thoracic casualties became considerably more conservative. Toward the end of the war, however, reports from general hospitals were to the effect that the incidence of infected hemothorax, empyema, and lung abscess was higher when foreign bodies were left in situ, and as a result, more thoracotomies began to be performed. In the 374 wounds of the chest cared for by the


5th Auxiliary Surgical Group, 142 thoracotomies were performed (38.00 percent), with 12 deaths (8.45 percent).

Hemothorax-Hemothorax was recorded as present in 296 of the 374 perforating or penetrating wounds (79.14 percent) and was stated to be absent in 24 cases. It was probably present in all the cases in which no statements were made about it. Hemothoraces varied in volume from 200 cc. to 2,500 cc. In the absence of other indications for interference, collections up to 300 cc. were left in situ. Larger collections were treated by aspiration, which was repeated as necessary. Air replacement was not employed, and no fluid was used except for a small amount of distilled water necessary to dissolve the penicillin (usually 40,000 units) used intrapleurally. If the chest continued to refill with blood after adequate aspiration, operation was considered to be indicated.

Active intrathoracic bleeding was found at thoracotomy in 50 wounds, 32 of which were caused by shell fragments. In 32 instances, the bleeding was from the lung (in 1 case from an intercostal vessel also), in 16 from an intercostal vessel, and in 2 from the internal mammary artery.

The case fatality rate was apparently related to the volume of the hemothorax. There were 2 deaths in the 49 cases in which the volume ranged up to 500 cc. and 12 in the 78 cases in which it ranged from 2,000 to 2,500 cc.

Tension pneumothorax-Tension pneumothorax occurred only 32 times, usually in valvelike sucking wounds. Treatment consisted of sealing the wound and emptying the pleural cavity by aspirations. If the collection of air was caused by an actual air leak from the lung, a large needle or stiff catheter was introduced into the pleural cavity via the second anterior interspace, and water-seal drainage was instituted.

Injuries of the lung-It was the practice to leave small lacerations of the lung unsutured unless they were bleeding. Larger lacerations were repaired.

In one instance, pneumonectomy had to be done for severe lacerations of both the upper and lower lobes. Recovery was uneventful. In another instance, massive intrabronchial hemorrhage occurred after suture of a large laceration and reexpansion of the lung by positive pressure. The patient died before the bleeding could be controlled. Autopsy showed that it came from the laceration and made it clear that a lobectomy should have been done. Both these injuries, however, are exceptions; valid indications for radical surgery for damage to the pulmonary parenchyma were extremely uncommon.

Cardiac wounds-Cardiac injuries were encountered 4 times in the 374 wounds. Two patients recovered uneventfully, one after suture of a laceration of the left ventricle and one after suture of a laceration of the right atrium and repair of a perforating wound of the lower lobe of the left lung. In the third case, autopsy revealed a shell fragment in the wall of the left ventricle, which had lacerated the descending ramus of the left coronary artery. Roentgenograms had indicated that the object was in the mediastinum. In the fourth case, autopsy revealed a shell fragment 25 by 15 by 10 mm. in the right ventricle. It had entered the superior vena cava and been swept down into the ventricle, lacerating the tricuspid valve in its passage.

Analysis of deaths-Of the 31 deaths in these 374 injuries, 6 occurred in gunshot wounds and 25 in shell-fragment wounds. Of the fatal wounds, 14 were on the right side and 14 on the left; 3 were bilateral. In one of the bilateral injuries, death occurred suddenly, 7 days after initial wound surgery, from a secondary hemorrhage from a lacerated azygos vein. Another patient developed bilateral pneumothorax and died of anoxia on the third postoperative day.

In 16 cases, death occurred on the operating table or during the first 12 hours after operation, chiefly from shock. In these 16 deaths were 4 caused by blast injuries and 1 caused by a skull fracture.

One fatality represented a serious but quite understandable error in diagnosis. The patient had sustained a penetrating, sucking wound of the left chest. According to the roentgenograms, the stomach seemed to be in the thorax. When, however, the patient


stated that as a boy he had had a left phrenicectomy for tuberculosis, the surgeon's assumption was that the diaphragm was elevated because of the previous operation and that the stomach was below the diaphragm, not above it. Surgery was therefore limited to debridement and closure of the chest wall, without intrathoracic exploration. The patient did not react well from operation and died on the third postoperative day. Autopsy revealed a congenital diaphragmatic hernia of the stomach, which had sustained a perforating wound.

Thoracoabdominal Wounds

Of the 165 casualties with thoracoabdominal wounds treated by the surgeons of the 5th Auxiliary Surgical Group, 4 died before operation could be started or were so obviously moribund that surgery was considered futile. On the other hand, the high case fatality rate from shock in the first 12-hour period after operation shows that surgery was not withheld merely because casualties were poor risks.

Basic data-Of the 165 thoracoabdominal wounds, 98 were on the left side, of which 27 were fatal; 67 were on the right side, of which 12 were fatal. The difference in the case fatality rates reflects the anatomic location and the vulnerability of the abdominal organs on the two sides.

When there were no thoracic injuries of consequence, the case fatality rates did not differ substantially from the rates in the total series (11 deaths in 62 injuries on the left, 6 deaths in 36 injuries on the right).

In most of the injuries (84 percent), the wound of entrance was in the thorax. In avulsing wounds and through-and-through wounds caused by small arms, it was often impossible to tell whether the course of the missile was from the thorax into the abdomen, or vice versa. In about a quarter of all cases, the thoracoabdominal wounds were complicated by other injuries.

Resuscitation-Many casualties with both thoracic and thoracoabdominal wounds were brilliant illustrations of what resuscitation can accomplish. One patient with a thoracoabdominal wound, whose blood pressure and pulse could not be obtained when he was first seen, received 17 units of plasma and 6 units of blood in the 10 hours before operation. He entered the operating room with a pulse of 110 and a blood pressure of 115/80 mm. Hg. 

On the other hand, shock was sometimes entirely irreversible. In these 165 patients with thoracoabdominal injuries, 47 needed intensive replacement therapy because of their poor condition and the severity of their wounds. Six of the forty-seven received 22 percent of all the blood and 30 percent of all the plasma given to this special group. They were finally deemed operable, but all six died of shock within the first 24 hours after operation. 

Pulmonary injuries.-There was no active bleeding in any of the 68 lacerations of the lung in these thoracoabdominal wounds. Suture repair was effected in 20 cases in which a bronchopleural fistula already existed or in which it was thought that one might develop. The other lacerations were left alone. The chest injury usually contributed only a small amount of the blood aspirated from the chest at operation. Most of it came from injuries of the spleen and, in small amounts, from injuries of the stomach and the small intestines.

Surgical approach.-There is no information as to the surgical approach in 25 cases in this series. In 39 wounds, 22 of which were on the left side, the approach was by the thorax. In 44 wounds, separate thoracotomy and laparotomy incisions were used, thoracotomy being the initial procedures in 33 cases and laparotomy in 11 cases. In the remaining 57 cases, 36 of which were on the left side, a laparotomy incision was used and no intrathoracic surgery was done.

If the wounding agent was a shell fragment, thoracotomy was the preferred mode of approach. The reasoning was that because of its larger size and its tumbling, rotating progress, such a missile was likely to create a gaping, sucking wound of the chest and diaphragm, whereas a diaphragmatic wound caused by small arms fire was likely to be sealed over. If the wounding agent was a bullet, the thoracic approach was preferred only


if there was evidence of a cardiac or vascular injury, a severe laceration of the lung, or herniation of abdominal organs through the diaphragm. Otherwise, a laparotomy was thought to be quicker and less shocking than the management of a bullet wound through a transdiaphragmatic incision, which would require enlargement of the diaphragmatic perforation to permit satisfactory exploration and repair.

As a rule, the transdiaphragmatic approach was used on the left side for penetrating wounds of the chest at the level of, or below, the seventh rib if the wound was caused by a projectile whose course was in an almost transverse plane. On the right side, the same indications were used if it was certain that the projectile had not proceeded to the region of the duodenum, right colon, or pancreas, all areas to which approach is blocked by the right lobe and main bulk of the liver. Wounds of the diaphragm located over the extraperitoneal surface of the liver could not be repaired by laparotomy alone; thoracotomy was required for adequate exposure.

The type of abdominal incision was governed by the course of the projectile and, to a lesser extent, by the degree of angulation of the costal margin at the xiphoid process. A subcostal incision in a patient with a wide costal flare provided excellent exposure for repair of the diaphragm and stomach and for exploration of the spleen and kidney, with splenectomy and nephrectomy if necessary. In the asthenic individual, a rectus incision provided good exposure and had the additional advantage of simplifying the placement of a stab wound for exteriorization of an injured colon.

Complications-In the 75 cases in which the liver was injured, there was no recognized instance of either hepatopleural fistula or bile empyema. Drainage in wounds of the liver was proportionate to the severity of the laceration. Volumes up to 1,500 cc. on the first day were not uncommon, and some wounds were still draining rather large amounts when the patients were evacuated.

Drainage from renal injuries was negligible, and it was concluded that a stab wound to provide for either urine or blood was probably unnecessary unless a major calyx of the renal pelvis was involved.

The chief complications in the 165 thoracoabdominal wounds were as follows:

Two bronchopleural fistulas developed within 24 hours of operation but closed spontaneously after institution of catheter drainage through the second interspace.

Two patients with atelectasis were successfully treated by bronchoscopy. Lesser degrees of atelectasis probably occurred and went undetected. Two patients had lobar pneumonia, in one case associated with an early empyema. This was the only instance of empyema observed before evacuation.

Wound infections were uncommon and usually not serious, but three wound disruptions occurred, one of which was fatal.

Paralytic ileus was present in most cases, sometimes for 4 to 6 days, in extensive injuries of the liver, retroperitoneal hematomas, or large hemoperitoneum.

Analysis of fatalities-Autopsies were performed in 21 of the 39 deaths in this series. The fatalities were distributed as follows:

Eighteen patients died of shock, fifteen within the first 24 hours after wounding. 

Eight patients died of acute pulmonary edema, five with a profuse terminal serohemorrhagic accumulation of fluid in the bronchial tree. In the four cases in which autopsy was performed, the fluid was found to originate in the lung, which was grossly enlarged, heavy, firm, and engorged with blood. Microscopic study revealed edema; acute passive congestion; and red blood cells in the alveoli, as well as areas suggestive of the consolidation of bronchopneumonia. In three of the four cases, the surgeon had no doubt that the cardiovascular system had been overloaded. One of these patients had received a unit of blood and 3 units of plasma before operation, and 4 units of blood and 11 of plasma during the operation. Another had received 1,500 cc. of physiologic salt solution and 2,000 cc. of blood in Alsever's solution over 2 hours.


Five patients died of pneumonia.

Three patients died of peritonitis, one with pneumonia.

Two deaths occurred from overlooked injuries. One patient succumbed to a rapidly spreading retroperitoneal infection from a missed perforation of the cecum. The other died on the fifth postoperative day, from hemorrhagic shock, after the clot in a missed perforation of the inferior vena cava became dislodged.

One patient with extensive laceration of the liver apparently died from biliary peritonitis.

One death was caused by the migration of a shell fragment, which was dislodged from the inferior vena cava and traveled to the right pulmonary artery, where it acted as an embolus.

One patient suffered an evisceration of a perforated stomach, small intestine, and transverse colon through a 7-by-6-cm. avulsed wound of the left lower chest laterally. After operation, his blood pressure ranged from 80 to 90 mm. Hg, and he was irrational and almost uncontrollable. His condition was presumably due to chronic shock. When renal shutdown was part of the picture, as it was in this case, no therapy was effective. 

Whether fat embolism played a part in any of these deaths is not possible to say. None of the microscopic specimens studied in the 21 autopsied cases were positive for fat.


General Considerations

The soundness of policies of early management of combat-incurred chest wounds could be evaluated by a study of the case fatality rates and morbidity rates of the casualties who survived to reach forward hospitals. Similarly, the soundness of policies of management in all oversea hospitals could be evaluated by an analysis of the residual status, and the necessity for further care, of the casualties received later in hospitals in the Zone of Interior. An analysis of the first 500 (of an ultimate 2,350) casualties with combat-incurred chest wounds, received in the thoracic surgery center at Kennedy General Hospital after it had become fully operational in June 1944, provides interesting and useful data of this kind.

Before these data are accepted absolutely, however, two qualifying statements are necessary:

1. It is only fair to emphasize that in a comparative analysis made retrospectively from field medical records and forward hospital records, it was not always easy, or indeed possible, to recognize all the circumstances that confronted surgeons working under combat conditions and that might have influenced their decisions concerning therapy.

2. The figures are frequently overlapping and, in a number of instances, though the records provided no actual proof, it was thought that the incidence of certain injuries and complications was probably higher than the stated figures indicated.

The material for this analysis was obtained by a careful review in each case of the field medical record, the records of previous hospitalizations overseas, the serial roentgenograms taken during the course of treatment, and the condition of the patient on his admission to the chest center. The clinical status of each patient was thus surveyed from the point of vantage of his latest hospitalization and in the light of his previous treatment. As a result, it was possible to make an objective evaluation of the character of the original injury, its initial management, the subsequent therapy, and the complications which followed special wounds and special techniques of management. The composite data permitted the comparative evaluation of surgical versus conservative therapy, forward thoracotomy and the relation of morbidity to its performance or omission, the use or omission of drainage in the closure of chest wounds, and similar routines.


Had a review of the last 500 patients admitted to the center been possible, there is no doubt that, as policies overseas had become stabilized in the light of experience, the outcome of the analysis would have been quite different.

Basic Data

All 500 wounds of the chest included in this survey were penetrating or perforating. Wounds limited to the chest wall very seldom reached thoracic surgery centers in the Zone of Interior. Small arms fire was responsible for only a few injuries, the great majority being caused by bombs and other high explosives.

The original injuries (the figures are sometimes overlapping) included:

211 retained foreign bodies of significant size.
136 sucking wounds.
75 thoracoabdominal wounds or combined thoracic and abdominal wounds. 
62 pulmonary lacerations.
45 bronchopleural fistulas. 
15 pulmonary contusions. 
9 tension pneumothoraces.
8 blast injuries.
4 fractures of the sternum. 
3 wounds of the esophagus. 
2 wounds of the trachea.
2 wounds of the pericardium.

As far as could be determined from the records, the chest wound was entirely uncomplicated in only 19 cases. Complications which occurred overseas included:

Hemothorax in 455 cases. In 151 of these cases, the hemothorax became infected, and in another 49, it became organized without infection.

Subphrenic abscess in 12 cases.
Pneumonitis and lung abscess about foreign bodies in 11 cases. The same complications developed in the 4 hematomas of the lung in the series.
Hemoptysis secondary to the presence of a foreign body in 8 cases. 
Atelectasis in 9 cases.
Pericarditis with effusion in 8 cases. 
Infected costal cartilage in 7 cases. 
Bronchopneumonia in 5 cases. 
Diaphragmatic hernia in 5 cases. 
Pulmonary infarction in 3 cases. 
Esophageal fistula and paraesophageal abscess in 3 cases. 
Pericardial tamponade in 2 cases.
Suppurative pericarditis in 2 cases.

These complications (again, the figures are sometimes overlapping) were managed as follows:

Drainage of empyema in 151 cases, in 46 of which secondary drainage was also necessary.
Thoracentesis in 47 cases. 
Decortication in 49 cases. 
Drainage of subphrenic abscess in 12 cases. 
Pericardial aspiration in 11 cases.
Closure of bronchopleural fistula in 9 cases. 
Repair of diaphragmatic hernia in 5 cases. 
Resection of ribs for osteochondritis in 4 cases.
Drainage of suppurative pericarditis in 2 cases and drainage to relieve pericardial tamponade in 1 case.


When these 500 patients were received at Kennedy General Hospital, the chest wound was healed in 314 cases (62.8 percent), and no further treatment other than rehabilitation was needed. The remaining 186 patients presented 213 complications, not all of which, as will be pointed out, developed from the original wound. Some of them were the result of the therapy employed. These complications included:

Retained foreign bodies in 131 cases.
Chronic empyema in 82 cases. 
Bronchopleural fistula in 36 cases. 
Draining sinus in 26 cases.
Hemothorax in 21 cases, in 12 of which the process was of the organized type. These figures should also be evaluated in comparison with the 82 cases of chronic empyema present when these casualties were received at Kennedy General Hospital.
Diaphragmatic hernia in 4 cases.
Esophageal fistula or paraesophageal abscess in 3 cases. 
Pericardial effusion in 2 cases.

Special Types of Injuries

Sucking wounds-The 136 sucking wounds listed in this series were of such a size as to require immediate closure when they were first seen in field or evacuation hospitals. A considerably larger number of patients had similar wounds, but they were so small as to be of no clinical significance. They responded promptly to packing or to debridement and primary closure. It should be noted, however, that infection developed in almost 10 percent of the cases in which the wound was closed primarily without drainage.

In larger sucking wounds, it was the practice to occlude the wound temporarily by packing and to close it permanently later, when debridement was done. Thoracotomy was done at this time in 78 of the 156 cases.

In a number of cases, the chest wound was only partly closed after debridement, and intercostal water-seal drainage was instituted. The records showed that convalescence was smoother in this group of patients than in any other patients with sucking wounds.

In several cases in which a bronchopleural fistula was present and had not been recognized, tension pneumothorax developed after closure of a sucking wound. 

Thoracoabdominal wounds.-In this series, 75 of the 500 patients had sustained injuries of the thorax and abdomen, either from a single missile that perforated the diaphragm or from separate perforations of the chest and abdomen. The abdominal organs injured were, in the order of frequency, the liver in 40 cases, the spleen in 16, the stomach in 15, the kidneys in 9, and the small intestine and colon in 7.

All these patients had been treated by prompt surgery, but the order of the procedure depended upon the severity of the chest wound. The chest injury was given first consideration in large sucking wounds, tension pneumothorax, pericardial tamponade, or rapidly developing hemothorax associated with shock and not relieved by conservative therapy. Otherwise, if the cardiorespiratory physiology was not seriously unbalanced, the primary attention was devoted to the abdomen.

The diagnosis of concomitant abdominal injury was frequently difficult when the missile had entered the chest, because in chest wounds in which there was no perforation of the diaphragm, abdominal pain, tenderness, and rigidity were often part of the picture. Urinalysis was always indicated in injuries of the lower chest, to determine whether or not blood was present. Intravenous pyelograms, which were made later in general hospitals, occasionally demonstrated defects in the renal pelvis secondary to injuries of the lower chest.

In the majority of cases in which the injury resulted from a single missile, exploration and the necessary surgery were successfully carried out through a transthoracic approach. When there was doubt concerning the extent and location of the abdominal injury, it was


FIGURE 1.-Posteroanterior roentgenogram showing large foreign body in contact with esophagus.

the general policy to use a separate abdominal approach because it permitted more extensive exploration.

The liver apparently tolerated the presence of metallic foreign bodies quite well. At any rate, no abscess resulted from failure to remove them, though biliary drainage followed the injury in a number of cases. When hepatic injuries were explored transpleurally, both the subphrenic space and the pleural space were usually drained. When this precaution was omitted, bile empyema, which occurred in 14 cases, and subphrenic abscess were both possibilities. The most serious instances of bile empyema occurred when the wound was closed without drainage.3

Intratracheal anesthesia was employed in all thoracoabdominal injuries, regardless of the surgical approach. The reason is obvious: If a tear of the diaphragm was present, atmospheric air entering from the abdominal incision could cause collapse of the lung on the affected side or even tension pneumothorax.

Tracheoesophageal injuries-There were only 3 injuries of the cervical esophagus and only 2 injuries of the upper end of the trachea in these 500 patients, but there were 3 similar injuries in the next 300 patients admitted to the Kennedy General Hospital chest center. The thoracic esophagus was not injured in any case in the series studied, even though, in a number of instances, ragged shell fragments were seen by roentgenograms to be lodged in the posterior mediastinum close to it (fig. 1). The high immediate case fatality rate resulting from the hemorrhage and the mediastinal infection commonly associated with injuries of this kind accounts for the small number of patients seen with them in Zone of Interior hospitals.

Three of the five patients with tracheal and esophageal injuries were submitted to immediate tracheotomy, and four had gastrostomies for feeding purposes within 4 to 10 days after wounding. All five developed fistulas of the esophagus associated with abscesses, which had to be drained. When the tracheotomy tubes were removed, from 2 to 3 weeks after operation, none of the patients had any difficulty in breathing.

Two patients with injuries of the esophagus developed strictures which required dilatation, and four of the five with combined injuries of the esophagus and the trachea developed persistent paralysis of the recurrent laryngeal nerve. Because these are extremely serious injuries, the eventual outcome in these cases was regarded as being as satisfactory as possible.

3Experience in the Mediterranean theater indicated that there was a significantly higher incidence of bile empyema when the diaphragm was closed with catgut rather than with silk.


FIGURE 2.-Posteroanterior roentgenogram showing Lipiodol visualization of bronchopleural fistula, narrow empyema track, and drainage site at some distance from fistula.

Bronchopleural fistulas-The records indicated that 45 of the 500 patients in this series had clinically significant bronchopleural fistulas immediately after sustaining their chest wounds. The chances are that fistulas were also present in a number of other cases but were so small that they required no special treatment and gave rise to no difficulties.

Thirty-six patients had bronchopleural fistulas when they were admitted to the chest center at Kennedy General Hospital, the majority secondary to their original injuries. Eight, however, followed the removal of foreign bodies, and six became evident only after the development of empyema. In other words, not all of these 36 fistulas were part of the original 45 fistulas.

All patients with fistulas observed at the chest center were studied by roentgenograms, after the introduction of iodized oil, to determine the size of the residual empyema space and the anatomic relation of the fistula to the site at which drainage had been instituted after thoracotomy (fig. 2). If the fistula was shown to be at some distance from the cutaneous wound and drainage would have been necessary through a long, narrow, empyematous track, healing was accomplished more rapidly if drainage was established directly over the fistula. When secondary drainage was instituted, exploration often showed that healing had been prevented by the presence of necrotic spicules of rib, bits of clothing, and other foreign bodies embedded in the pulmonary tissue. The removal of these objects and the excision of scar tissue about the fistula were usually followed by prompt healing, though in an occasional case, these measures did not suffice and closure of the fistula had to be accomplished by means of a muscle flap.

Retained Foreign Bodies

As the incidence of immediate and delayed complications shows, the major problems in the management of injuries of the chest concerned retained foreign bodies; hemothorax and pneumothorax; empyema; and the use of thoracotomy, with or without drainage, in forward areas.


FIGURE 3.-Posteroanterior roentgenogram showing pneumonitis and fibrosis associated with retained foreign bodies.

Management overseas-Of the 211 patients observed with intrathoracic foreign bodies of considerable size when they were first examined after wounding, 47 underwent removal of the objects at the primary operation and 33, after an interval of several weeks. Many other fragments were noted in other patients when roentgenologic examination was carried out, but their size and location were such as to make them of no clinical significance. 

Throughout the war, differences of opinion continued to exist as to the clinical significance of metallic foreign bodies and the indications for their removal. The chief reason for removing them was the possible development of hemorrhage or infection, as well as other complications, but the incidence of these manifestations in this series was not great. In the 164 cases in this series in which the objects were not removed at initial wound surgery, hemoptysis occurred later in 8, draining sinuses developed in 8, lung abscesses in 7, and pneumonitis in 6. In other words, there were no untoward manifestations in 135 of the 164 cases in which the foreign bodies were not removed at initial wound surgery. Moreover, practically all of these complications developed within 2 to 4 weeks after wounding.

In many cases in this series, the removal of the foreign bodies at the initial operation was part of the procedure of debridement, hemostasis, and repair of lacerated lung tissue. In this group, the removal was an incidental step in the operation.

In other instances, however, initial wound surgery was undertaken primarily for the removal of the foreign body. The objects were often hard to locate in newly traumatized tissue, and fresh trauma was created. Bronchopleural fistulas were sometimes created and went unrecognized. If the injury was small, the chest was frequently closed without drainage. As a result of these various circumstances, the incidence of postoperative infection in this group of patients was almost 30 percent.

Generally speaking, infection was most frequent in the cases in which dirt, particles of clothing, and similar debris entered the chest with the foreign body. This extraneous


FIGURE 4.-Foreign bodies in contact with large blood vessels. A. Posteroanterior roentgenogram. B. Same.

material was more likely to be the cause of infection than the metallic object itself. Foreign bodies unaccompanied by other debris were usually found at operation, as previous roentgenologic evidence had shown, to be associated with surprisingly little reaction and with no infection. Objects in contact with large blood vessels were practically always well encapsulated in fibrous tissue.

Management in the Zone of Interior-Of the 131 patients who entered Kennedy General Hospital with roentgenologic evidence of metallic foreign bodies still in situ, only 49 harbored missiles considered of clinical significance because of their size, shape, location, or clinical manifestations. Only 23 of these 49 objects were removed, on the following indications:

1. Large, irregular shell fragments lying within the lung parenchyma were removed without hesitation, particularly if there was an associated pneumonitis or fibrosis (fig. 3). 

2. Similarly, no hesitation was felt about removing foreign bodies lying in the mediastinum in contact with the esophagus (fig. 1) or with large blood vessels (fig. 4).

3. Foreign bodies associated with lung abscesses (fig. 5) or empyema were removed at the time the necessary drainage was instituted. Four patients with retained foreign bodies had either pneumonitis or lung abscesses when they were admitted to the center. All four had draining sinuses, and three complained of hemoptysis.

4. In other cases, foreign bodies causing hemoptysis or intercostal pain were reproved whenever it could be demonstrated that they were directly responsible for these symptoms (fig. 6). In most such cases, the onset of symptoms occurred shortly after wounding. There was no evidence of infection or other reaction in any case in this series in which roentgenologic examination showed no abnormalities for 2 to 3 months after wounding. This was not, of course, an infallible rule: In one of these cases gross hemoptysis appeared a year after injury.4

5. In an occasional case in which operation was not otherwise indicated, the object was removed because of the psychologic factors of fear or pain on the part of the patient.

4Attention is called to the patient observed in the Mediterranean theater who required pneumonectomy for a retained foreign body 15 years after wounding.


FIGURE 5.-Retained foreign bodies giving rise to infection. A. Posteroanterior roentgenogram showing lung abscess associated with retained object. At operation, clothing and other debris were found in the abscess cavity. B. Posteroanterior roentgenogram showing foreign body partly embedded in lung parenchyma and protruding into empyema cavity.

As the war progressed, the policy at the Kennedy General Hospital thoracic surgery center became increasingly conservative, and foreign bodies were removed only on strict indications. The postwar experience will have to determine the late possibilities for harm of retained foreign bodies and thus establish clear-cut criteria for their removal. It is unfortunate that plans for a formal followup of men with retained shell fragments never came to fruition (vol. I).

Tetanus toxoid was always given before operations for the removal of foreign bodies. Otherwise, the preoperative and postoperative regimen, including the use of penicillin, was the same as for other thoracic operations.


Management overseas-Hemothorax was recorded after wounding in 455 of the 500 patients in this series, 91 percent. In many cases, it was present in association with other conditions of serious import, such as sucking wounds, lacerations of the lung tissue, and retained foreign bodies.

The most severely injured patients in this group were submitted to thoracotomy in frontline hospitals, on entirely justifiable indications. In many other cases, however, casualties whose injuries were considerably less severe were also submitted to thoracotomy, on indications that were far from clear cut. The effectiveness of the method of treatment employed in each case is best determined by the presence or absence of infection. The figures are as follows:

Of 74 cases treated expectantly, without any positive therapy, 30 became infected (40.54 percent). In 34 cases, the hemothoraces were small and spontaneous healing occurred. Infection occurred in 6 of the 12 cases of bronchopleural fistula.

Of 156 cases treated by thoracotomy, 75 (48.08 percent) became infected (figs. 7, 8, and 9).


FIGURE 6.-Retained foreign bodies giving rise to symptoms. A. Lateral roentgenogram showing metallic foreign bodies responsible for hemorrhage 1 year after wounding. B. Posteroanterior roentgenogram showing retained metallic foreign body responsible for intercostal pain.

FIGURE 7.-Early thoracotomy without drainage in penetrating bullet wound of chest. Immediate thoracotomy was performed for hemostasis and to remove the bullet, which was found so located in the mediastinum that it could not be removed. The chest was closed without drainage. A. Posteroanterior roentgenogram showing postoperative hemothorax. Note bullet in left apex. The hemothorax later became infected and required open thoracotomy drainage. B. Posteroanterior roentgenogram showing total empyema space.


FIGURE 8.-Early thoracotomy without drainage in penetrating shell-fragment wound of right chest. A. Posteroanterior roentgenogram showing retained shell fragment and hemothorax. B. Same, after operation, showing pneumothorax secondary to bronchopleural fistula. The wound was closed without drainage, and the resulting empyema had to be drained on three occasions.

FIGURE 9.-Early thoracotomy without drainage in perforating wound of right chest. A. Posteroanterior roentgenogram showing preoperative hemothorax. B. Same, 2 days after thoracotomy without drainage, which was omitted because the amount of contamination was thought to be minimal. C. Same, showing fluid level in right hemithorax indicative of postoperative empyema.


Of 225 cases treated by repeated thoracenteses, only 38 (16.89 percent) became infected. If the 13 cases are excluded in which aspiration was inadequate and only the 212 properly treated patients are considered, the proportion of infection falls to 11.8 percent. In 11 cases in this group, the chances of good results were reduced by the associated bronchopleural fistulas.

Of the 49 patients whose hemothoraces became organized, 20 had either had no aspiration of the chest or only a single aspiration, and 7 had been aspirated only twice. Nine had been treated by thoracotomy. All of these patients underwent decortication overseas.

The comparative figures leave no doubt of the soundness of thoracentesis as the preferred method of treatment in cases in which immediate operation is not indicated for some reason other than the presence of a hemothorax. The circumstances of military medicine were always in favor of the simplest method of treatment, in this instance thoracentesis, as compared to more complicated methods, in this instance thoracotomy. The necessity for evacuation, the changes of medical personnel from echelon to echelon, and the complicated care required in handling thoracotomy drainage all help to explain the higher incidence of infection when thoracotomy was used in preference to thoracentesis.

Management in the Zone of Interior-All the cases in which decortication was done at Kennedy General Hospital further confirmed the advantages of this operation (fig. 10). It proved a logical means of developing an easily obtainable cleavage plane between the fibrous encasement of the lung and the adjacent viscera, pleura, pericardium, diaphragm, and chest wall. When the peel was separated from these structures, reexpansion of the lung occurred promptly, the lung tissue herniating through the line of incision as soon as separation was completed.

The experience at this center also confirmed the general opinion that the best results were obtained when the operation was done between 4 and 6 weeks after wounding. If the lung was allowed to remain collapsed for 3 months or more, some fibrosis of the parenchyma frequently took place, and reexpansion after operation was delayed.

There is no doubt that decortication, as it was practiced in World War II, was responsible for many of the good results obtained in casualties with chest wounds (fig. 11) who, in the past, would have become invalids from chronic fibrothorax.


Management overseas-Of the 151 patients who developed empyema in association with their chest injuries, 82 had chronic empyema when they were admitted to the chest center. In 62 of the 151 cases (41.06 percent), the cause of the empyema could be traced directly to the circumstances of the original injury, such as its initial severity, the extensive contamination introduced by the missile, an associated laceration of the lung, or the presence of a bronchopleural fistula.

In the remaining 89 cases, the development of the empyema could be traced wholly or in part to two causes, (1) failure to carry out adequate aspiration of a hemothorax when no other surgery was necessary, and (2) failure to provide postoperative water-seal drainage when surgery had been done.

No patient who had water-seal drainage developed an empyema of any clinical significance, but infection resulted in 31.4 percent of the patients who had been submitted to thoracotomy without postoperative drainage and in 18 percent of those who had had inadequate thoracenteses after operations in which surgical drainage was omitted (figs. 7, 8, and 9). In another 9.4 percent of the cases, the thoracotomy tube had been removed too soon. The analysis of these cases thus suggests that the incidence of empyema in combat-incurred chest injuries could be appreciably reduced if thoracentesis were performed adequately after operation and if adequate water-seal drainage were provided in all cases.


FIGURE 10.-Observations at decortication. A. Organized hemothorax after evacuation of fibrinous mass from pleural space. Note organized fibrin on visceral pleura. B. Easily obtainable plane of cleavage between fibrous encasement and visceral pleura. C. Reexpansion of lung by positive pressure anesthesia after removal of fibrous encasement.

Drainage was carried out overseas in all 151 cases of empyema, on an average of 3 weeks after the infection was detected. Secondary drainage was necessary in 46 cases. In the first months of the war, there was a decided tendency to perform open thoracotomy for infected hemothorax, early in the illness, and total empyema seems to have followed this mode of treatment rather frequently (fig. 12). Later, this policy was almost entirely discontinued, and drainage was deferred until this risk no longer existed.


FIGURE 11.-Infected organized hemothorax managed by decortication. A. Posteroanterior roentgenogram before operation. B. Same, showing results of decortication.

Management in the Zone of Interior-The majority of patients admitted to the chest center with chronic empyema had been adequately drained, and healing was progressing well. If drainage was not adequate, secondary drainage was instituted. Most of the cavities were either small or moderate-size.

Adjunct treatment of chronic empyema usually consisted of continuous suction with 30 to 50 cm. of negative pressure. Breathing exercises and blow bottles were also employed, but the experience was that reexpansion of the lung was more apt to occur, and to occur at a faster rate, when suction was employed. Dakinization of the wound was also employed in a few cases. Other constituents of therapy included repeated transfusions of blood and plasma, high-vitamin diets supplemented by the administration of vitamins by other routes, and general hygienic measures.

As time passed, it was found that empyema secondary to hemothorax could be treated by decortication with practically the same degree of success as could be secured in organized hemothorax without infection. When this method of treatment came into use, it was only an occasional patient with empyema who had to be submitted to thoracoplasty.

Thoracotomy in Forward Hospitals

Many of the cases in this series proved, as has already been intimated, that the mere presence of a penetrating or a perforating wound of the chest did not, in itself, constitute an indication for thoracotomy in a forward hospital. Many operations, in fact, designed to stop hemorrhage or to prevent infection were themselves followed by these and other complications, particularly when the chest was closed without drainage.

Of the 500 patients in this series, 156 were submitted to debridement, with thoracotomy and repair of injured pulmonary tissue, within 24 to 48 hours after wounding. Of these, 6 percent developed organized hemothoraces, and 76 (48.7 percent) developed infection. These patients fared much worse than the 225 patients with hemothorax treated by thoracentesis, in only 16.8 percent of whom was additional treatment necessary because of infection.


FIGURE 12.-Infected hemothorax and total empyema. A. Posteroanterior roentgenogram showing early infected hemothorax with mediastinal shift. B. Same, showing total residual empyema after early open thoracotomy drainage. C. Same, showing healed chest after continuous suction therapy over 4-month period.

It must be granted that the surgical group included the severest injuries, for which prompt surgery was mandatory and in which complications might have developed if operation had not been done. Furthermore, it was not always easy to determine from records alone the circumstances which led frontline surgeons to follow the courses which they did. Nonetheless, in 66 of the 156 cases in which early thoracotomy was done, the field medical records show no clear-cut indications for the procedure. The majority of the patients in this group presented both hemothoraces and retained foreign bodies, but neither can be considered a valid indication for thoracotomy, and there seems to be no doubt, from the subsequent incidence of infection after the operation, that they would have fared at least as well, and probably better, if they had been treated conservatively.


Since chemotherapy and antibiotic therapy were employed routinely in both the group treated conservatively and the surgical group, a comparison of results from this standpoint would shed no light on the incidence of infection in the two groups. An analysis of the use or omission of drainage in the surgical group, however, produces some useful data:

In addition to the 156 patients submitted to thoracotomy as a primary procedure, 33 other patients underwent the same operation at a later time overseas. In these 189 cases, drainage was instituted in 81 and omitted in 108. In 67 of the 81 drained cases (82.7 percent), healing occurred smoothly, and there was no need for secondary surgery. The patients who became infected were, for the most part, those with the most severe injuries. 

In contrast, in 61 of the 108 cases in which drainage was omitted (56.5 percent), secondary operations were required. Furthermore, healing in the undrained group covered an average of 14 weeks, while the average in the drained group was 10 weeks.

The higher incidence of postoperative infection, the greater necessity for secondary surgery, and the longer convalescence all make clear that a thoracotomy closed without drainage achieved desirable results in a much smaller proportion of cases than a thoracotomy supplemented by adequate drainage. The contrast was particularly evident in the 16 patients who were found at primary thoracotomy to have lacerated lung tissue. When the wounds were closed without drainage and operation was not followed by adequate thoracentesis, many of these patients required secondary drainage for infection (fig. 7).

Other considerations also suggested that routine water-seal drainage after thoracotomy in frontline hospitals was a far safer plan than closure without drainage. The conditions under which a surgeon had to care for the wounded in the combat zone usually did not permit him to supervise postoperative care except for a brief period. It was therefore not possible for him to supervise personally the thorough aspiration of a hemothorax, which was essential in all undrained chest cases. Another practical consideration was that waterseal drainage required considerably less time and attention than repeated thoracenteses. This was an important matter in a busy frontline hospital.

As these cases indicated, transportation was not a contraindication to water-seal drainage. While the patient was en route to the rear, the thoracotomy tube could be clamped and the bottle temporarily disconnected, or a flapper type of drain could be used, as was done in a number of these cases in which results were excellent.


An analysis of the first 500 chest injuries received at the thoracic surgery center at Kennedy General Hospital indicated that the oversea management of these cases was generally good. In a Zone of Interior chest center, as might have been expected, mortality is minimal; there were no deaths in this series. Morbidity, however, is a matter of consequence. This study suggests that it can be reduced by the following policies:

1. Thoracentesis should be substituted for operation in the management of hemothorax unless other indications for early surgery exist.

2. Thoracotomy in forward hospitals should be limited to such clear-cut indications as sucking chest wounds, tension pneumothorax, and hemorrhage. The mere existence of a chest wound is not an indication for forward surgery.

3. Thoracentesis should be performed adequately after operation, or adequate waterseal drainage should be instituted.