CHAPTER XVII
Wounds of the Stomach (416 Casualties)
Luther H. Wolff, M. D.
An analysis of the 416 wounds of the stomach (table 38) whichoccurred in the 3,154 abdominal injuries treated by the 2d Auxiliary SurgicalGroup during 1944 and 1945 makes three points clear:
1. Wounds of this organ were considerably more frequent than they hadheretofore appeared to be.
2. They were complicated by injuries to other viscera in 9 out of every 10cases.
3. The case fatality rate in gastric injuries (40.6 percent)was significantly higher than it was in injuries of the colon, small intestine,liver, spleen, or genitourinary tract (q. v.).
TABLE 38.-Essential data in 416 wounds of stomach
Type of wound | Cases |
| Deaths | Case fatality rate | |
| In 416 gastric injuries | ||||
|
|
| Percent |
|
|
Univisceral | 42 | 1.3 | 10.1 | 12 | 28.6 |
Multivisceral | 374 | 11.9 | 89.9 | 157 | 42.0 |
| 416 | 13.2 | 100.0 | 169 | 40.6 |
One hundred and ninety-six of these gastric injuries (47.1 percent) wereproduced by missiles which traversed the diaphragm. There were 85 deaths inthese thoracoabdominal injuries (43.4 percent). In the remaining 220 cases, theprojectiles entered or traversed only the abdominal cavity. There were 84 deathsin this group (38.2 percent).
In addition to the unexpectedly high frequency of injuries of the stomach inrelation to total injuries in this series, the ratio of univisceral tomultivisceral wounds is at marked variance with data reported in other series ofabdominal wounds (table 39). The frequency was nearly twice as great as thatpreviously reported. Actually, this is what might be expected. As has beenpointed out (p. 92), the incidence of wounding of any organ is almost directlyproportional to the space which it occupies. It follows, therefore, since thestomach
224
is a relatively large organ, that it would be injured fairlyfrequently. Furthermore, the frequency of wounding was probably even higher thanis indicated by these figures, which take no account of casualties who diedbefore reaching a field hospital.
Source | Total series |
| Multivisceral wounds of stomach | ||
Cases |
| Cases | Frequency (wounds of stomach) | ||
|
|
| Percent |
|
|
World War I: |
|
|
|
|
|
| American | (2) | 144 | 7.0 | 48 | 33.3 |
| British | 965 | 82 | 8.5 | 26 | 31.7 |
Spanish Civil War | 238 | 20 | 8.4 | (2) | (2) |
World War II: |
|
|
|
|
|
| British3 | 628 | 29 | 4.6 | 15 | 51.7 |
| 2d Auxiliary Surgical Group | 3,154 | 416 | 13.2 | 374 | 89.9 |
1The American World War I figures are from the MedicalDepartment of the United States Army in the World War (Washington: U. S.Government Printing Office, 1927, vol. XI, pt. 1, pp. 65, 457, 458). The BritishWorld War I figures were reported by Cuthbert Wallace in War Surgery of theAbdomen (London: J. & A. Churchill, 1918). The Spanish Civil War series wasreported by Douglas W. Jolly in Field Surgery in Total War (New York: Paul B.Hoeber, Inc., 1941). The British World War II figures were reported by W. H.Ogilvie in Abdominal Wounds in the Western Desert (Surg., Gynec. & Obst.,March 1944).
2Not stated.
3Western Desert combined series.
The marked predominance of multivisceral over univisceralwounds is also what might be expected. Univisceral wounds are naturallyinfrequent in an organ almost completely invested by the liver, spleen, colon,and kidneys. It is believed that the figures for this series more nearlyapproach both the true frequency of wounds of the stomach and the true ratio ofunivisceral and multivisceral gastric injuries than any statistics previouslyreported, if only because in World War II more casualties survived to undergosurgery than in any previous war.
DIAGNOSTIC CONSIDERATIONS
Experience showed that only two signs could be regarded as conclusive in thepreoperative diagnosis of a wound of the stomach. One was the emission ofundigested food from an abdominal wound. The other was the actual observation ofa perforation or laceration in an eviscerated stomach. In the absence of thesetwo signs, the diagnosis was only presumptive.
According to the literature, vomiting is a cardinal sign of awound of the stomach. An analysis of these 416 gastric wounds does not supportthis observation. Vomiting was no more frequent in wounds of the stomach than
225
in wounds of other abdominal viscera. It was actuallyrecorded only 7 times in the 416 cases, and inquiry of resuscitation officers andof surgeons in the group revealed that none of them considered it an outstandingor a reliable sign of gastric injury.
Blood in the vomitus or in the aspirated gastric contents was, however,regarded as both suggestive and reliable. It was noted altogether in 41 woundsof the stomach (4 times in vomitus), and constituted a valuable clue to thenature of the injury when blood swallowed in wounds of the head, neck, or lungscould be ruled out. The absence of blood was not recorded consistently, beingmentioned only eight times. The negative observation, even when it was recorded,was obviously of no diagnostic value.
The preoperative passage of a Levin tube, in the opinion ofthe surgeons of the group, not only did no harm in suspected wounds of thestomach but was, on the contrary, a useful diagnostic and therapeutic measure.The risk of introducing possible contamination to the injured site was faroutweighed by the relief of the accumulations of gas and fluid which were commonin gastric and other abdominal injuries and whose persistence led to increasedleakage of gastric contents and more serious peritoneal contamination.
Leakage of gas from the damaged stomach was sometimes avaluable diagnostic sign, though it could also produce a variety of confusingbut interesting clinical pictures. Thus subcutaneous emphysema of varyingdegrees was at times observed in the abdominal and chest walls, and in someinstances gas actually bubbled from the abdominal wound. If the gastric woundwere thoracoabdominal, gas from the stomach might escape into the pleural cavitythrough the lacerated diaphragm and produce pneumothorax.
Pneumoperitoneum was a matter of record in only six of the roentgenogramstaken in wounds of the stomach. This is a deceptively low figure, since 90percent of all casualties with abdominal wounds had preoperative roentgenographic examinations. The circumstances, however, were not conducive to preciseroentgenography. All the examinations were made with portable apparatus.Practically always, as a concession to pain and shock, the patients were notmoved from the supine position. Lateral views were impractical and were seldomattempted. Finally, conditions for both development and interpretation of thefilms were difficult. Under the circumstances, there is no doubt that a gasbubble lying free in the peritoneal cavity was sometimes overlooked.
In some cases, as soon as the peritoneum was opened, thesurgeon was greeted by a rather disconcerting gush of air. It was oftendifficult to determine whether it originated in the stomach or from the chest,through a perforation of the diaphragm, and careful exploration was necessaryto determine its origin. The presence of gas in the abdomen also introducedstill another diagnostic problem-the possibility of an anaerobic infection.Crepitus and discoloration of tissues might be the result of gas and leakage ofacid secretions from a perforation of the stomach, but it might also beexplained by an early infection of this kind.
226
The preoperative diagnosis of a wound of the stomach dependedchiefly on visualization of the course of the missile and the application ofaccurate anatomic knowledge concerning the location of the organ. In the greatmajority of these wounds, the location of the wounds of entry and exit-if theinjury was perforating-and the location of the wound of entrance, combined withlocalization of the missile by two-plane roentgenography-if the injury waspenetrating-permitted an accurate preoperative diagnosis.
Variations in the shape of the stomach and the position ofthe soldier at the time of wounding (p. 100) naturally complicated thediagnosis, as the following case history indicates:
Case 1.-A prisoner of war was seen in a field hospital with a wound ofentrance in the left hip just above the head of the femur and a wound of exit inthe right hip through the wing of the ilium. A low midline exploratory incisiondisclosed multiple perforations of the small bowel and sigmoid colon. It also,rather unexpectedly, disclosed a severe laceration of the stomach some 4inches above the upper end of the midline incision. Reconstruction of theinjury suggested that the stomach was in the lower abdomen at the time ofwounding, the man undoubtedly being crouched over in the position a combatsoldier automatically assumes when he is under fire.
NATURE OF THE INJURY
Wounds of the stomach varied widely in type (table 40). In 16 cases, thewound was a simple tangential laceration of the stomach wall, without penetration into the lumen. At the other extreme were 5 cases in which the violenceof the trauma resulted in complete transection of the organ. In the remainingcases, the wound varied from a trivial perforation to a laceration 20 cm. long.
A fair proportion of the wounds were caused by small missiles, which perforatedone or both walls of the stomach in a nearly perpendicular plane. Perforationsof this kind often caused little or no peritoneal contamination from gastricleakage, since the redundant gastric mucosa tended to act as a valve and sealoff the injury. Shock was not usually severe. A fairly large number of suchinjuries accounted for the relatively low case fatality rate in perforatingwounds of the stomach (table 40). In contrast, missiles which entered the wallof the stomach at an acute angle were likely, regardless of their size, toproduce
TABLE 40.- Casefatality rates in relation to type of injury in 416 wounds of stomach
| Type of injury | Cases | Deaths | Case fatality rate |
| Perforating | 258 | 91 | 35. 3 |
| Laceratingl | 117 | 71 | 60. 7 |
| Not stated | 41 | 7 | 17. 1 |
| 416 | 169 | 40. 6 |
1Includes 5 complete transections.
227
extensive lacerations. Leakage was profuse, peritonealcontamination was severe, and the case fatality rate was correspondingly higher(table 40). Larger missiles tended to produce lacerating wounds, which werepotentially more lethal, and part of the difference in the case fatality ratesof perforating and lacerating wounds can be explained on this basis.
TIMELAG AND THE MULTIPLICITY FACTOR
As has already been mentioned (p. 217), the timelag was ofthe utmost importance in individual wounds of the viscera. In the whole seriesof 416 injuries, however, it did not seem of serious significance unless theanalysis also took into account such other considerations as the multiplicity ofvisceral injuries, the number and character of associated injuries. and thedegree of peritoneal contamination.
Univisceral wounds of the stomach (table 41) were so few (42, 10.1 percent)that their analysis from the standpoint of the timelag would be of nostatistical importance. This is, also as already noted, a considerably lowerfrequency than is reported in any other comparable series. The case fatalityrate, 28.6 percent,was surprisingly high. Although the multiplicity factor (p. 111) was found to bea more reliable prognostic index in abdominal injuries than any other factor,the general rule that the greater the number of organs injured, the higher thecase fatality rate (fig. 21,p. 108) did not hold in wounds of the stomach in thefirst two multiplicity categories (table 41). It is true that in a category ofonly 42 univisceralwounds, statistical error is likely. On the other hand, an analysis of the 12 deaths in these 42 injuries seems to provide a clue tothe lethality of a type of wound which in other structures was attended with aproportionately lower case fatality rate. This analysis showed that 7 of the 12 deaths in univisceral gastric injuries occurred eitheron the day of operation or within the first 2 days after operation, and that the cause was recordedin every instance as shock or shock and peritonitis. In two other cases, deathoccurred from peritonitis, one on the fourth and one on the eighth postoperativeday. Another death was caused by peritonitis and a gastric fistula on the 15thpostoperative
TABLE 41.-Influence ofmultiplicity factor on case fatality rates in 416 wounds of stomach
Organs injured | Cases |
| Case fatality rate |
Stomach only | 42 | 12 | 28.6 |
Stomach and 1 viscus | 174 | 47 | 27.0 |
Stomach and 2 viscera | 112 | 44 | 39.3 |
Stomach and 3 viscera | 50 | 29 | 58.0 |
Stomach and 4 viscera | 23 | 23 | 100.0 |
Stomach, other viscera, and great vessels | 15 | 14 | 93.3 |
| 416 | 169 | 40.6 |
228
day. One of the two remaining deaths was caused by secondaryhemorrhage from the stomach, on the 14th postoperative day. In the other case,no cause of death was stated. These are small figures and of no statisticalsignificance, but a clinician cannot fail to be impressed by the causes and timeof death and by the number attributed to shock.
In multivisceral injuries (table 42), the liver, as might be expected, wasmost frequently involved, with the colon and the spleen next in order.Concomitant injury to the colon produced the highest case fatality rate exceptfor concomitant injuries to the great vessels. The latter were almostuniversally fatal (table 41).
One extremely significant fact in the analysis of thesemultivisceral wounds was the discovery that treatment had been successful incertain combinations of wounds which up to this time had been uniformly fatal.Bailey,1 for instance, stated in 1942 that in the past no combination ofwounds involving the stomach, small intestine, and colon had ever beensuccessfully treated. In the 24 injuries of this kind in this series, in 11 ofwhich the liver was also injured, there were 15 survivals.
TABLE 42.-Case fatalityrates in wounds of stomach complicated by wounds of other viscera1
Organs injured | Cases |
| Case fatality rate |
Stomach only | 42 | 12 | 28.6 |
Stomach and liver | 67 | 20 | 29.9 |
Stomach and spleen | 42 | 8 | 19.0 |
Stomach and colon | 24 | 11 | 45.8 |
Stomach, colon, and liver | 18 | 9 | 50.0 |
Stomach, liver, and spleen | 17 | 5 | 29.4 |
Stomach and jejunum | 16 | 4 | 25.0 |
Stomach, jejunum, and colon | 13 | 4 | 30.8 |
Stomach, jejunum, colon, and liver | 11 | 5 | 45.5 |
Stomach and kidney | 10 | 3 | 30.0 |
Stomach, jejunum, and liver | 9 | 3 | 33.3 |
Stomach and ileum | 7 | 0 | 0 |
Stomach, colon, and spleen | 6 | 2 | 33.3 |
Stomach, jejunum, and kidney | 6 | 2 | 33.3 |
Stomach, spleen, and kidney | 6 | 2 | 33.3 |
Stomach and pancreas | 6 | 1 | 16.7 |
Stomach, colon, and kidney | 5 | 5 | 100.0 |
Stomach, colon, liver and spleen | 5 | 2 | 40.0 |
Stomach, liver, and pancreas | 5 | 1 | 20.0 |
1This tabledoes not include combinations of visceral wounds which occurred less than 5times each. It also does not include 2 injuries of the stomach and duodenum inwhich both patients survived and 2 injuries of the stomach and great vessels inwhich both patients died.
1Bailey, Hamilton: Surgery ofModern Warfare. Edinburgh: E. & S. Livingstone, 1941-42.
229
ASSOCIATED INJURIES
About a quarter of the 416 casualties with wounds of the stomach also had severeextra-abdominal injuries. Forty-three presented major fractures, forty-one majorsoft-tissue injuries, nine major amputations, nine injuries of the spinal cord,four injuries of the brain, and four injuries of the heart. The chest injuriespresent in thoracoabdominal wounds were, strictly speaking, separate associatedlesions, and their presence apparently produced a 5-percent increase in the casefatality rate in this group. An evaluation of the effect of the associatedinjury and its influence on morbidity and case fatality rates would require acase-by-case analysis. Generally speaking, the rate among patients withassociated injuries did not differ significantly from that of patients withoutthem, though the fallacy of the application of such a generalization to theindividual case is obvious.
TECHNICAL CONSIDERATIONS
The transdiaphragmatic approach to wounds of the fundus and body of thestomach (table 43) greatlyfacilitated the repair of injuries in these areas, both the anterior andposterior surfaces of the fundus being accessible through the incision withoutopening of the peritoneal folds. It is, therefore, not surprising that thisincision was used in 119 (60 percent) of the 196 thoracoabdominal injuries. Eventually, it became theincision of choice in selected cases, though it was used in only one instance inwhich there was not a perforation of the diaphragm. In that type of injury, avertical incision in the upper abdomen was preferred.
The gastrocolic omentum was incised routinely at operation, to permitexamination of the posterior gastric wall. Whether operation was done from aboveor from below the diaphragm, this was a most important step of the procedure inpatients known or suspected to have sustained gastric injuries.
TABLE 43.-Distribution of surgical approaches in 412 wounds of stomach1
| Cases | Percentage |
Laparotomy | 293 | 71.1 |
Transdiaphragmatic: |
|
|
| Thoracotomy | 95 | 23.1 |
| Combined laparotomy and thoracotomy | 18 | 4.4 |
| Thoracolaparotomy2 | 6 | 1.4 |
| 119 | 28.9 |
| 412 | 100.0 |
1Information is lacking on this point in fourcases.
2By this technique, the thoracic incision isextended across the costal arch onto the abdominal wall.
230
Simple suture, which was used in 409 of the 416 injuries of the stomach, wasregarded as the procedure of choice, even in extensive lacerations. Resectionwas required in the five cases in which transection of the stomach was complete.The Polya or Hofmeister technique was used in three cases, all of which werefatal, and end-to-end anastomosis in the other two, one of which was fatal.
Perforations of the stomach were overlooked at operation in the two remainingcases in the series. One of the patients died of a gastropleural fistula; theother died of embolism. At autopsy, the overlooked perforations were discovered;they had closed spontaneously, without apparent leakage, and had not contributedto the fatal outcome.
The type of suture closure, as well as the kind of suturematerial, varied with the preference of the individual surgeon. Except thatpurse-string suture proved unsatisfactory, as will be pointed out shortly,variations in technique seemed to make no difference in the outcome of the case.
In six cases, all simple perforations, severe postoperative hemorrhageoccurred from the stomach. This was the largest number of postoperativehemorrhages encountered in the entire series of 3,154 abdominal injuries. Threepatients died. Two survived under conservative management, and the otherrecovered uneventfully after a secondary operation to control bleeding.
It issignificant that in all six instances of postoperative bleeding, closure hadbeen by reenforced purse-string suture. This method can produce circumstancesideal for the development of hemorrhage. Since the suture seldom, if ever,includes the gastric mucosa, the mucosal edges retract. As a result, the bloodvessels which traverse the submucosa are exposed, and the resulting pathologicpicture simulates that of acute peptic ulceration. Erosion of the previouslysealed underlying vessels, and subsequent hemorrhage, are thereforepossibilities whenever this method is used. The following case historyillustrates this fact:
Case 2.-A soldier with multiple wounds from shell fragments, including a leftthoracoabdominal wound, was admitted to a field hospital in excellent generalcondition. Transdiaphragmatic repair of a single perforation of the upperportion of the stomach was carried out by means of a purse-string suture througha left thoracotomy incision. The operation was performed without complications,and convalescence was smooth until the fifth postoperative day, when evidencesof internal bleeding were observed. When laparotomy was resorted to, after 7hours of conservative therapy, the stomach was found to be completely occupiedby a clot estimated to contain 1,500 cc. of blood. After the clot had beenremoved, the site of perforation and suture was inspected from the mucosal side.The surgeon's notes adequately explained the hemorrhage: "A white induratedarea is seen from which the mucosa is retracted. From the edges, in two places,are seen continuous but small streams of blood, one venous and the otherarterial. This ulcer-like area, then, is the cause of all bleeding."Excision of the affected area and closure of the wound were followed by anuneventful recovery.
This report and others like it led to certain technical changes:
1. Every effort was made to approximate the gastric mucosa by suture in allwounds of the stomach.
231
2. Small perforations were enlarged transversely so that the mucosal layercould be exposed and accurately sutured.
3. Purse-string suture was no longeremployed in wounds of the stomach.
POSTOPERATIVE COMPLICATIONS
As already noted, practically every patient who died withinthe first 48 hours after operationas the result of a wound of the stomach presented the clinical picturedesignated under the generic term "shock." All had sustained extensivetissue damage, massive blood loss, disturbed pulmonary physiology, peritonealcontamination of varying degrees, or combinations of these conditions.Persistent shock resulting from these conditions is, strictly speaking, apostoperative complication, but it is not classified as such in this discussionbecause of the time of its occurrence. Actually, it was part of the originalpathologic process.
Of the complications which occurred later than 48 hours after operation, by far the commonest werepulmonary. In fact, pneumonia (12 cases), empyema (11 cases), and atelectasis (6cases) accounted for approximately 40 percentof the serious complications seen after this time. Peritonitis was observed in 6cases, 5 of which were fatal. In all instances, it was of the usual clinical(bacterial) type.
CASE FATALITY RATE
The data analyzed in this section bear out the initialstatement that wounds of the stomach, whether univisceral or multivisceral, areamong the most serious injuries encountered in warfare (table 44). Some perforations were comparatively trivial, but theleakage of acid gastric contents into the general peritoneal cavity was alwaysserious and was likely to be associated with a high degree of shock. In general,patients with wounds of the stomach exhibited a much more severe degree of shockthan those with other abdominal wounds. In 22 cases (13 percent of thefatalities) death occurred on the operating table; death occurred in the courseof operation in only 11 percent of all other fatalities in the total series ofabdominal injuries. The 115 fatalities (68 percentof the total deaths from wounds of the stomach) which occurred by the end of thesecond postoperative day were chiefly attributable to shock alone or inassociation with peritonitis and, to a lesser degree, other complications, or tooverwhelming peritoneal contamination.
In 12 of the 416 casualtieswith wounds of the stomach, the records contributed nothing concerning thepresence or absence of shock. In 66 othercases, the patients were either in no shock at all or in incipient shock. Onehundred and three of the remaining patients were in mild shock, one hundred andsix in moderate shock, and one hundred and twenty-nine in severe shock. Theunusually high proportion of casualties in moderate and severe shock can beexplained, at least in part, by the spillage of acid gastric contents into the
232
general peritoneal cavity. There seems no doubt that wheneverspillage of such contents occurred, an immediate chemical peritonitis ensued,which quickly produced either shock or a shocklike state and which was entirelydifferent from the bacterial peritonitis observed in later stages of the injuryin patients who survived.
TABLE 44.- Primarycause of death in relation to time of death in 169 wounds of stomach
Time and cause of death |
|
On operating table: |
|
| Shock | 13 |
| Shock and hemorrhage | 5 |
| Shock and atelactasis | 1 |
| Shock and gas gangrene | 1 |
| Shock and peritonitis (contamination) | 1 |
| Cardiac | 1 |
| 22 |
Through second postoperative day: |
|
| Shock | 43 |
| Shock and peritonitis (contamination) | 28 |
| Shock and hemorrhage | 4 |
| Shock and atelactasis | 3 |
| Shock and anaerobic infection | 2 |
| Peritonitis (massive contamination) | 9 |
| Peritonitis (massive contamination) and pneumonia | 2 |
| Peritonitis (massive contamination) and intestinal fistula | 1 |
| Pulmonary embolism | 1 |
| 93 |
After second postoperative day: |
|
| Peritonitis (bacterial) | 5 |
| Peritonitis (bacterial) and shock (all third day) | 8 |
| Peritonitis (bacterial) and pneumonia, empyema, or pleurisy | 12 |
| Anuria | 10 |
| Pneumonia | 6 |
| Hemorrhage, secondary | 3 |
| Pneumothorax and pleurisy | 3 |
| Intestinal or gastric fistula | 2 |
| Brain injury | 1 |
| Not stated | 4 |
| 54 |
Another possible explanation of shock in wounds of thestomach originates in anatomic considerations, that the stomach overlies ahighly vascularized area in which are located the celiac axis, the aorta, theinferior vena cava, and
233
the portal vein, and that massive hemorrhage is therefore frequentlyassociated. Although this is theoretically sound reasoning, hemorrhage ingastric wounds did not appear to be more severe than in numerous other visceralwounds in this series. The part which it played in the production of shock wasoften difficult to evaluate, it is true, but in the usual wound of the stomachit seemed to be secondary in importance to peritoneal contamination.
After the second postoperative day, peritonitis and pulmonary complicationswere the most important causes of death (table 44). The case fatality rate amongcasualties with lacerating wounds was almost double that among those withperforating wounds (table 40), the explanation probably being that all patientswith lacerating wounds suffered massive peritoneal flooding with acid stomachcontents, which presumably happened in a much smaller proportion of those withperforating wounds. The high case fatality rate among patients with univisceralwounds of the stomach (table 41) is further proof of this hypothesis. Althoughthe relative vascularity of the stomach and the adjacent structurestheoretically should play a part in the lethality of gastric wounds, thecollective clinical observations of the 2d Auxiliary Surgical Group did notsupport this point of view.
