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 21



Wounds of the Liver and of the Extrahepatic Biliary Tract 
(829 Casualties)

Gordon F. Madding, M. D., Knowles B. Lawrence, M. D., and Paul A. Kennedy, M. D.

The 829 wounds of the liver which are discussed in this chapter occurred in 3,0661 of the 3,154 abdominal injuries observed by the 2d Auxiliary Surgical Group during 1944 and 1945 (table 73). Four hundred and forty-six (53.8 percent) were thoracoabdominal. The case fatality rate in this group was 24.0 percent (107), against 27.0 percent (224) for the whole group. In 51 of the 829 wounds of the liver, the gallbladder was injured, and the extrahepatic biliary ducts were injured in 2 others. There were 16 deaths in these 53 cases.

TABLE 73.-Essential data in 829 wounds of liver

Type of wound




Case fatality rate

In 3,066 abdominal injuries

In 829 liver injuries


























The majority of these injuries occurred in the right lobe of the liver, as might have been expected because of its greater size. In the 538 cases observed in 1944 in which these data were accurately stated, the right lobe was involved 446 times (82.9 percent), the left lobe 73 times, and both lobes 19 times. In-

1Calculations for injuries of the liver are made on a total of 3,066 abdominal injuries, instead of on 3,154, the base figure for all other calculations. When the 88 histories which represent the difference between the two totals became available, the medical officers who had done the work on wounds of the liver had been transferred from the command, and it was thought wiser not to change their tabulations.


volvement of adjacent viscera could usually be explained by the location of the hepatic injury. Wounds about the hilum were seldom seen, possibly because its protected location kept it from injury but more probably because most such injuries were immediately fatal. Injuries in this area involved such structures as the extrahepatic biliary ducts, the retroperitoneal portion of the duodenum, the pancreas, the stomach, the colon, and the vena cava.

Most wounds of the liver were lacerating, penetrating, or perforating. Penetrating wounds were usually less serious than those which involved larger masses of tissue, though they were sometimes followed by graver complications, such as bile leakage or hemorrhage. Twenty-four percent of all hepatic injuries were severe, sometimes severe enough to require resection of a part of a lobe. The remaining wounds were described as of slight or moderate severity.

Bleeding from the liver, which was only occasionally described as severe, had ceased at the time of exploration in 91.1 percent of all cases. The significance of this observation will be discussed later (p. 278).

Some bile leakage had occurred in practically every case, but the presence of blood and exudate in the peritoneal cavity, and sometimes of intestinal content also, made an estimate of the amount of bile spill impossible.


The management of wounds of the liver by surgeons of the 2d Auxiliary Surgical Group was a matter of evolution and the result of their own clinical observations. They had no decisive principles of therapy to guide them. Observations of military surgeons in World War I were brief and frequently contradictory. Bailey2 advocated expectant treatment if it were possible to be certain that the wound affected only the liver. Lee3 and Wallace4 also favored nonintervention in most cases. The Committee on Surgery of the National Research Council5 apparently felt that, in World War II, operative therapy would be necessary more often than it had been in the past.

All of these observers recognized the possibility that spontaneous hemostasis might occur in injuries of the liver, but they placed different degrees of reliance upon it. Bailey felt that physiologic changes incidental to anesthesia and surgery might cause renewed bleeding. On the contrary, Wallace's6 experience that spontaneous hemostasis usually occurred within 6 to 10 hours of wounding is in close agreement with the experience of the 2d Auxiliary Surgical Group.

The observers cited were in general agreement about the possibility of secondary hemorrhage from the liver, although there was some variance of

2Bailey, Hamilton: Surgery of Modern Warfare. Edinburgh: E. & S. Livingstone, 1941-42.
3The Medical Department of the United States Army in the World War. Washington: U. S. Government Printing Office, 1927, vol. XI, pt. 1.
4Wallace, Cuthbert: War Surgery of the Abdomen. London: J. & A. Churchill, 1918.
5Abdominal and Genito-Urinary Injuries. Military Surgical Manuals. Prepared under the auspices of the Committee on Surgery of the Division of Medical Sciences of the National Research Council. Philadelphia and London: W. B. Saunders Co., 1942.
6See footnote 4.


opinion about its frequency and danger. Makins7 stated that secondary hemorrhage did not occur in the absence of sepsis.

Lee8 advised suture of the liver in preference to packing. Wallace9 gave diametrically opposite advice. Bailey,10 while admitting the theoretic superiority of suture, nevertheless concluded that packing was "infinitely better" because of the technical difficulties attending suture of the liver. Packing, suture, and free muscle grafts were all mentioned by the Committee on Surgery of the National Research Council as permissible methods for the accomplishment of hemostasis and repair of hepatic wounds.11 Comments on the use of simple drainage and on the risk of bile peritonitis were notably lacking in these discussions of hepatic injuries.

The greatest measure of agreement among surgeons prior to World War II was their advocacy of expectant or conservative treatment, or no treatment at all, for most wounds of the liver. Many of these men were surgeons who had carried over the idea from their experience in World War I. This policy, however, was incompatible with the surgical philosophy of World War II, in which operative intervention was held to be in the best interests of practically every casualty with an abdominal wound if he lived to reach a forward hospital. Since the liver was frequently wounded, both alone and in combination with other organs, the problem of management of hepatic injuries became important early in the war.

During the initial activities of the 2d Auxiliary Surgical Group in 1942-43 (a period not covered by these data), wounds of the liver were variously managed by packing, suture, free muscle grafts, and expectant treatment. Suture and muscle grafts proved impractical in the hands of most surgeons, and expectant treatment was not thought to be reliable. During this period, therefore, gauze packing was the most frequently used of all methods. As time passed, however, serious complications were observed following this technique. The removal of gauze packs in field hospitals was more than once attended by disastrous hemorrhage. Abscesses occurred within the liver or the perihepatic space, and necrosis of hepatic tissue occurred in areas which had been packed. Hepatitis and biliary peritonitis were other complications. These and other undesirable results of gauze packs were emphasized in Burford's12 report of hepatic injuries observed in a general hospital.

Early removal of the pack was practiced in some cases in an attempt to prevent these complications, but this plan was not successful. As soon as the pack was removed, the external wound tended to close prematurely. Attempts to substitute Penrose drains for packs were also unsuccessful. Adequate

7Makins. Cited by Bailey. See footnote 2, p. 276. 
8See footnote 3, p. 276.
9See footnote 4, p. 276. 
10See footnote 2, p. 276. 
11See footnote 5, p. 276.
12Review of 1,000 Thoracic Cases. Bull. U. S. Army. M. Dept. 89: 70-72, June 1945.


drainage by this means proved impossible except under direct vision, when the abdomen was open.

By early 1944, many surgeons of the group had concluded that the use of packs in wounds of the liver was associated with too many risks to justify their employment if it could possibly be avoided. They had become aware of the frequency with which spontaneous hemostasis was observed in fresh wounds, and it seemed to them that the only justification for the use of a pack was the prevention of secondary hemorrhage. Since this was not a frequent complication, a trend began away from the use of gauze packs and toward simple drainage of the subhepatic space (Morison's pouch). At first there was some disagreement concerning the necessity for drainage, but the presence of bile in discharges in a substantial proportion of cases when both drains and gauze packs were used, as well as the frequent observation of bile in the peritoneal cavity in wounds of the liver, convinced most surgeons of the necessity for external drainage as a protection against bile peritonitis.

Some surgeons continued to close the abdomen without drainage when the injuries of the liver were limited to lacerations or penetrations which required no special treatment. The majority, however, took the position that all injuries required drainage, on the ground that it was impossible to predict, in a given case, whether bile leakage would or would not occur. In a few patients with thoracoabdominal wounds, troublesome bile empyema resulted from failure to drain the subphrenic space after operation had been performed through the right thorax. It was the consensus that this complication would probably not have occurred in these cases if drainage had been instituted through a subcostal incision.

These clinical observations led to almost complete abandonment of packing as a method of treatment in hepatic injuries and to its almost universal replacement by drainage of Morison's pouch. The data recorded in this chapter substantiate the soundness of this change of policy. The almost total absence of postoperative hemorrhage in the cases which were drained indicates that this risk had been overemphasized early in the war. Similarly, the efficacy of drainage as a means of preventing bile peritonitis is attested by the absence of this complication in the cases treated by this method.


The location of the incision in injuries of the liver (table 74) naturally depended, in any given case, upon the extent of the injury or injuries, and, in particular, upon whether the wound was abdominal or thoracoabdominal. The popularity of the transdiaphragmatic approach increased as the war progressed. In an occasional case, thoracolaparotomy was employed; the incision was extended over the chest wall onto the abdomen, the costal arch being cut or the incision passing through the 10th or 11th intercostal space.

In many wounds of the liver, there was sufficient fragmentation of the organ to necessitate actual resection of the damaged tissue instead of debride-


ment. Often a substantial amount of tissue had to be removed. In this sort of case, it was not unusual to find completely detached pieces of liver lying free in the peritoneal cavity as a result of the original trauma.

TABLE 74.-Distribution of surgical approaches in 432 wounds of liver
































Combined laparotomy and thoracotomy





















1In this technique, the thoracic incision was extended across the costal arch onto the abdominal wall.

In a sample of 695 cases observed in 1944 and 1945 in which full data concerning treatment were available (table 75), the changing proportions of cases treated by drainage and by packing are a significant reflection of increasing surgical experience. Though originally some wounds of the liver were regarded as too small to require drainage, this attitude became less and less frequent as the war progressed. In some cases in which drainage was instituted, there was no bile in the discharge after operation, but, as already noted, it was impossible, either preoperatively or at operation, to identify the cases in which this could be expected to happen. The size of the missile was not the determining consideration. A wound caused by a small foreign body which cut a main bile channel was likely to be followed by more profuse drainage than a superficial hepatic injury of greater extent. The principle that all suspected wounds of the liver should be explored and the additional principle that all should be treated by adequate drainage were eventually established as the only sound policy in this type of injury.

TABLE 75.-Percentage distribution of operations in 695 wounds of liver





Drainage only




Suture and drainage




Suture and pack




Pack only




Muscle graft




No treatment





Drainage was instituted with Penrose or cigarette drains. Two were employed if the wound was large or if there was more than one injury. The accepted practice was to place a Penrose drain lateral to the postero-inferior margin of the liver, to prevent collections of bile or pus in this region, and to use the same technique for drainage of the subhepatic space. To guard against wound infections and disruptions, the drains were not brought to the exterior through the original laparotomy or thoracolaparotomy incision. Instead, they were delivered through a dependent incision placed subcostally in the anterior or midaxillary line and preferably at least 1.5 inches long. It was found essential to extend this incision through all layers of the abdominal wall. If they were not widely opened, the drains were likely to become strangulated and their whole purpose would be defeated. In the occasional case, a debrided wound track coincided with the site of the usual subcostal incision for drainage and could be used satisfactorily for this purpose.

It was imperative to keep the drains in situ until bile drainage had ceased completely. This was usually by the 10th or 12th day, though it was sometimes considerably later. Drains were always shortened gradually, beginning on the 4th or 5th day. They were never removed abruptly in toto.

Although packing became less and less popular as a definitive method of treatment as the war progressed, the temporary use of a dry pack was occasionally extremely helpful. Active oozing from large superficial wounds of the liver could be controlled by this means while more urgent injuries were being cared for. It was unusual, when the pack was removed, not to find that bleeding had ceased and that no other measures were necessary to control it.

Relatively few wounds of the liver were sutured by surgeons of the group in the course of the war. By 1945, most of them had abandoned the practice completely, chiefly because it had been repeatedly observed that more serious bleeding might follow this procedure than had existed before it.


Not very many complications of wounds of the liver were observed in forward hospitals as compared with the number observed in hospitals farther to the rear. A report by Burford13 from a general hospital showed that 25, chiefly resulting from inadequate drainage, had occurred in 98 wounds of the liver. They included subphrenic abscess in 14 cases, intrahepatic abscess in 6 cases, and bile empyema in 5 cases.

The complications observed in field hospitals were what might have been expected in view of the nature of the wounds. They were chiefly presented as bile empyema or a biliary thoracic fistula. The diaphragm was obviously perforated in all thoracoabdominal wounds, which comprised 53.8 percent of the wounds of the liver. Bile and exuded fluids, if they could not drain externally because a pack had been used or for other reasons, sometimes forced their way through the sutured diaphragmatic wound, with the results just

13See footnote 12, p. 277.


stated. The technique by which the diaphragm was closed seemed to have nothing to do with their occurrence. In one such case, the lung was adherent to the diaphragmatic suture line, and the erosion of a bronchus by bile created an extremely serious problem. Occasionally the bile, after it had eroded the diaphragmatic suture line and reached the pleura, caused a breakdown of the thoracotomy incision, with a resulting subphrenic pleurocutaneous fistula.

Although there was no instance in a forward hospital of a fatal secondary hemorrhage, a surgeon of the group witnessed such an accident in a general hospital after removal of a pack from the liver on the 18th postoperative day. The single instance of serious postoperative bleeding which occurred while these patients were under observation in forward hospitals is recorded in detail because of the lessons implicit in the case:

Case report-An American soldier was admitted to a field hospital after being wounded in the right costophrenic sulcus and the hilar region of the liver by high-explosive shell fragments. The wound was extended and debrided under nitrous oxide-oxygen-ether anesthesia, and a shell fragment and several bits of clothing were removed from the liver. There was some oozing of dark blood, but the bleeding was not thought sufficient to justify any special hemostatic measures. Penrose drains were introduced and brought out through a separate drainage incision.

For 36 hours after operation, there was copious drainage of bile, but the postoperative course was otherwise uneventful until the eighth postoperative day. Then a severe hemorrhage occurred through the drainage incision. It ceased spontaneously, and there was no further bleeding until the ninth day, when a second severe hemorrhage occurred. Like the preceding hemorrhage, it was massive and obviously arterial in origin. It was checked by the insertion of a gauze pack deeply into the liver wound, through the enlarged drainage incision. A third hemorrhage 36 hours later was checked by the insertion of a fresh pack after the removal of the original pack.

Chills and fever occurred on the 16th postoperative day, while the second pack was still in situ. On the 17th day, the temperature rose to 105.8? F., and another severe hemorrhage occurred about the pack. The external wound was then opened under Pentothal anesthesia, and the wound in the liver was exposed by excision with the actual cautery for a depth of 7 cm. A large artery was found to have been partially severed by the original injury. It was clamped and ligated, and Penrose drains were inserted into the depth of the wound.

Convalescence was stormy. Severe distention developed on the 18th day (counting from the day of the first operation). On the 22d day, there was profuse biliary drainage. On the 32d day, a liver abscess was drained. On the 41st day, a pelvic abscess was drained, and on the 51st day a subhepatic abscess was drained. Thereafter recovery was satisfactory, and when the patient was evacuated on the 75th day, he was in good condition.

The hemorrhages in this case were probably all caused by repeated reopening of the partially severed artery as the surrounding clot retracted. The insertion of the first pack, on the ninth day after operation, was perhaps justified, in an attempt to control bleeding by conservative means. In retrospect, however, it clearly would have been wiser to explore the bleeding area without delay when hemorrhage recurred; it was known that the wound was in the hilum, and the hemorrhage was so massive that extensive packing was necessary to check it. There was no local or systemic evidence of infection when this pack was applied. It seems certain, however, that the compli-


cations which followed-infection, secondary venous bleeding, hepatic abscess, subhepatic abscess, and pelvic abscess-all arose because a pack was used and drainage was thus impeded. Any one of these complications could readily have been fatal, and all might have been prevented had the secondary hemorrhage been attacked at its source when it first occurred on the eighth postoperative day.


The case fatality rate in these 829 wounds of the liver, 27.0 percent (224 deaths), is to be compared with the rate of 66.3 percent reported for similar wounds in American Expeditionary Forces in World War I.14 These 829 wounds represented 27.0 percent of 3,066 abdominal injuries, in contrast to the 13.3 percent represented by comparable wounds among American soldiers in World War I (table 8, p. 93). The British proportion of wounds of the liver in that war was 16.8 percent (table 8).

Shock was responsible for 115 of the 224 fatalities in wounds of the liver, 51.3 percent of the total number of deaths. All deaths from this cause occurred before the end of the second postoperative day. When these patients were first seen, they were all in serious circulatory collapse, which persisted in spite of vigorous efforts at resuscitation. Blood loss was only one of several contributory factors; others included disturbances of cardiorespiratory physiology, overwhelming contamination of the peritoneal and pleural cavities, actual destruction of tissue, and widespread retroperitoneal cellulitis.

Pulmonary complications were the cause of 38 deaths, 17.0 percent of the total number. Trauma to the diaphragm, which occurred in over half of all wounds of the liver; trauma to the lung; contamination of the pleural cavities by bile; and the necessary prolongation of anesthesia and operation provided an ample background for this cause of death.

Peritonitis was primarily responsible for 28 deaths, 12.5 percent of the total number. It was present to some degree in all fatal cases, but it was not considered the primary cause of death unless it was widespread or took the form of a localized process, such as a subphrenic abscess.

Oliguria and renal failure were the chief causes of death in 19 cases (8.5 percent). The so-called transfusion kidney and pigment nephropathy could not be excluded as the principal lesion in these cases, but interesting possibilities obviously suggest themselves concerning the relationship between the liver damage and the renal failure.

In the remaining 24 cases, the causes of death included such conditions as gas gangrene, head injuries, and wounds of the spinal cord with paraplegia. The hepatic injury in this group usually played only a minor role in the fatal outcome.

It should be emphasized again that in no case in this series could death during the postoperative period in a forward hospital be ascribed to bleeding

14See footnote 3, p. 276.


from the liver. In World War I, hemorrhage was the chief source of both morbidity and mortality in wounds of the liver, while in World War II, at least in the cases handled by the 2d Auxiliary Surgical Group, bile leakage and damage of the hepatic parenchyma (after shock) were the chief factors of mortality.

The source of bleeding in wounds of the liver is the hepatic artery, the hepatic vein, or the portal vein. Unless, therefore, a hilar injury has been sustained or a missile has penetrated deeply into the liver tissue, serious bleeding should not be common. The larger branches of the portal vein lie nearer the surface than the branches of the arterial system, but the pressure in the venous system is low (8 to 10 mm. Hg), and hemorrhage from them, as is clear from this series of wounds of the liver, is not difficult to control.

The change in therapeutic methods reflects this change of emphasis. In World War I, packs to control hemorrhage furnished the chief method of treatment when active treatment was undertaken at all. In World War II, active therapy was undertaken in all cases, and drainage eventually became the most popular method of treatment.

The death rate in this series undoubtedly was influenced by certain factors such as the wounding agent (p. 97), the timelag from injury to operation (p. 103), the availability of adequate shock therapy (p. 124), and the use of the sulfonamides and later of penicillin (p. 197). It is difficult, however, to reduce their influence to statistical terms. The case fatality rate was higher during the winter than the summer, probably, as in other injuries, because of the higher winter incidence of respiratory infections.

The case fatality rate was directly proportional to the number of viscera injured (table 10, fig. 22). It was 9.7 percent when only the liver was injured, but it rose to 84.6 percent when four or more other viscera were injured also.

Associated extra-abdominal injuries were chiefly compound fractures of the long bones, traumatic amputations, injuries of the head and spinal cord, and injuries of the lungs and other thoracic structures, including, occasionally, the heart. They added greatly to the morbidity of this series and undoubtedly contributed to the death rate, though it is impossible to evaluate their individual influence.


The degree of damage to the gallbladder varied widely in the 51 wounds of the liver in which this organ was implicated. In one instance, the fundus was partially avulsed from its bed; but the gallbladder wall was not damaged, and simple suture was the only treatment necessary. In five cases, the wounds of the fundus were small and could be closed by purse-string suture. In the remaining cases, the damage was almost evenly divided between severe lacerations which required cholecystectomy and less severe injuries which could be treated by cholecystostomy.


One of the two patients with an injury of the common duct suffered a perforation of the hepaticoduodenal ligament; simple suture, without drainage, was followed by prompt recovery. The other had a wound of the duct near the ampulla of Vater. It was overlooked at operation, and, while the patient had other severe injuries, this error unquestionably contributed to the fatal outcome.

Because of the presence of wounds of the liver in all wounds of the extrahepatic biliary tract, it is clearly impossible to determine the influence exerted by the latter type of injury upon the case fatality rate. It was probably not very important. The rate for all wounds of the liver was 27.0 percent. For the 53 cases in which wounds of the liver were associated with injuries of the extrahepatic biliary tract, it was 30.2 percent, which is not materially different.