Acute Gastric Disease After Cutaneous Thermal
Injury
MAJ Albert J. Czaja, MC, USA; MAJ Joseph C. McAlhany, Jr., MC, USA; MAJ Willard A. Andes, MC, USA; COL Basil A. Pruitt, Jr., MC, USA
Diffuse erosive "gastritis" was discovered as early as five hours postinjury in 45 of 54 burn patients (83.5%) evaluated by gastroduodenoscopy. Acute ulcers were identified in 14 patients (26%); concomitant duodenal disease was present in 34 patients (76%). Microvascular fibrin thrombi were not demonstrated even though five patients had disseminated intravascular coagulation. Seven patients were examined before nasogastric intubation; four, with a mean burn size of 59.6% total body surface, had diffuse "gastritis." Low total serum protein levels were measured in 81% of tested patients, but were not predictive of mucosal disease. Hemorrhage followed the clinical deterioration of six patients (11.1%); one ulcer perforated. Whereas coagulation abnormalities, nasogastric intubation, and hypoproteinemia may augment mucosal injury, the morphologic and histologic examinations of the lesions suggested a primary ischemic cause resulting from the opening of submucosal shunts or local vasoconstriction.
The development following
of acute erosions of the gastric mu¬ cutaneous thermal injury was first described by Dupuytren in 1832.' Retrospective clinical and postmortem studies2'5 have recognized this erosive "gastritis" as a life-threatening complication in the burn population, occurring with a frequency estimated between 14.4%- and 24.7%."' Endoscopie examinations of the stom¬ ach and duodenum within the immediate postburn period have indicated an even higher incidence of disease." Acute mucosal ischemia secondary to the opening of submucosal cosa
Accepted
a
for publication Dec 15, 1974. From the US Army Institute of Surgical Research, Brooke Army Medical Center, Fort Sam Houston, Tex. Read before the 82nd annual meeting of the Western Surgical Association, San Francisco, Nov 22, 1974. The opinions or assertions contained herein are the private views of the authors and are not to be considered as official or as reflecting the views of the Department of the Army or the Department of Defense. Reprint requests to US Army Institute of Surgical Research, Brooke Army Medical Center, Fort Sam Houston, TX 78234 (Dr. Czaja).
arteriovenous shunts,1-7 8 local vasoconstriction,71 or the de¬ velopment of microvascular thromboses from a dis¬ seminated intravascular coagulation1"" or postburn hemoconcentration1- has been implicated in the cause. An indwelling nasogastric tube (routinely used to decompress the burn patient with ileus), hypoproteinemia, and a nega¬ tive nitrogen balance may augment the mucosal injury.13 To determine the clinical and pathologic characteristics of this acute gastric disease, early and serial endoscopie ex¬ aminations of the stomach and duodenum were performed in thermally injured patients admitted to the US Army Institute of Surgical Research. In 29 patients examined within 72 hours postburn, coagulation studies and mucosal biopsy specimens, specially stained to detect fibrin thrombi, assessed the etiologic importance of microvascular thromboses. The interrelationship between the acute gastric disease, the indwelling nasogastric tube, and hypoproteinemia was also evaluated. METHODS All nonpediatric patients admitted within one week after sus¬ taining a burn of more than 25% of their total body surface were eligible for endoscopy. A history of chronic gastrointestinal symp¬ toms or excessive alcohol or aspirin ingestion eliminated the pa¬ tient from the protocol. Each patient was managed independently by his physician. The informed written consent of the patient or his next of kin was obtained prior to each procedure. Endoscopie examination of the entire stomach and proximal part of the duodenum was usually accomplished within 72 hours postburn, utilizing a fiberoptic panendoscope. This evaluation was repeated during the second and the third week of hospitalization. Other ex¬ aminations were performed whenever indicated clinically. Intra¬ venous administration of diazepam (up to 20 mg), occasionally supplemented by intravenous administration of meperidine hy¬ drochloride (up to 50 mg), was begun immediately before each en¬ doscopie procedure. Photographs documented the mucosal disease.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Table 1.—Correlation of Acute Gastric Disease With Burn Size, Duodenal Disease, and Clinical
% of Total
Body Surface
No. of Patients
Burn 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 90-100 Total
Incidence,
Erosive Gastritis*
Acute Gastric
Ulcerf
Acute Duodenal Disease
Gastric
Hemorrhaget
Complications
Gastric Perforation
Operations for Gastric Disease
0
14 16
12 16
54
45 83.5
/o
* Fundus and body always involved; antrum also diseased t Multiple gastric ulcers in five patients (36%). X Bright red bleeding via nasogastric tube.
14
14 26 in 15
6 11.1
34 63
1 1.86
1 1.86
patients (33%).
Biopsy specimens were obtained from the gastric lesions and the adjacent intact mucosa. These specimens were processed with hematoxylin-eosin and phosphotungstic acid-hematoxylin stains; histologie sections were examined for the presence of microvascular fibrin thrombi. The platelet count, fibrinogen level, pro¬ thrombin time, partial thromboplastin time, and other selected co¬ agulation studies, including fibrin degradation products, serial thrombin time, plasminogen, and factors V and VIII, were deter¬ mined on admission and were reevaluated every two to four days during the first two weeks of hospitalization. These studies were reviewed by a hematologist who formulated a clinical inter¬ pretation of the coagulation changes while unaware of the endo¬ scopie findings. A sudden, marked decrease in both the platelet count and fibrinogen level in conjunction with a prolongation of the prothrombin and partial thromboplastin times suggested the diagnosis of disseminated intravascular coagulation. Serum albu¬ min and total protein levels were monitored during this same interval. Patients with gastrointestinal ileus were routinely decom¬ pressed with a No. 18 F nasogastric sump tube attached to inter¬ mittent low pressure suction. Endoscopy was performed immedi¬ ately before or shortly after nasogastric intubation. Occasionally, the sump tube was in place for up to 72 hours before the initial en¬ doscopie examination. Localized or linear mucosal lesions were at¬ tributed to intubation trauma.
RESULTS
Seventy-nine endoscopie examinations were performed without complication in 54 patients. All patients were male, with three exceptions. Ages ranged from 16 to 76 years (mean age, 33.4 years) and burn sizes ranged from 20% to 96% of the total body surface (mean burn size, 55.6% of the total body surface). Mortality within this study group was 76% (41 of 54 patients). Postmortem ex¬ amination in 36 cases (88% of the patients studied) cor¬ roborated the findings of gastroscopy. Twenty-nine patients were evaluated within 72 hours postburn; one patient was examined as early as five hours postinjury. Twenty-five other patients were studied from four to 14 days after their burn. Mucosal biopsy specimens were obtained from 13 of the patients evaluated within the initial 72-hour period, and coagulation studies were
completed in 24. Serum protein levels were measured dur¬ ing this same period in 21 patients. Acute
Superficial Gastric Mucosal Disease ("Gastritis") and Ulcération
Table 1 indicates the incidence of erosive "gastritis" and acute ulcération, and the relationship of acute gastric disease to burn size, duodenal disease, and clinical compli¬ cations. Mucosal lesions were not discovered in patients with burns involving less than 30% of their body surface. As indicated in Table 2, acute gastric mucosal disease was present in 76% of patients evaluated within 72 hours postinjury. Erythematous petechial lesions, confluent macular areas of mucosal hemorrhage, and punctate erosions on a pale mucosa were identified within the fundus and body of the stomach (Fig 1). The antrum was also involved in 15 of the 45 patients with gastric disease (33%). Concomitant duodenal disease was present in 34 of the patients with "gastritis" (76%). Acute gastric ulcers were not discovered before 72 hours postburn and always occurred in areas of severe superficial mucosal disease. Five of the gastric ul¬ cer patients (36%) had multiple craters. The histologie examination of the macular lesions showed microvascular congestion, edema, and focal muco¬ sal hemorrhage. Superficial epithelial necrosis above the muscularis mucosae, usually with an exúdate of fibrin and neutrophils, characterized the erosions. Mucosal regenera¬ tive activity was occasionally evident (Fig 2). Frank mucosal infarction was also encountered (Fig 3). Mucosal necrosis extending beyond the muscularis mucosae distin¬ guished the ulcers. Inflammatory reaction around the ul¬ cer was only mild to moderate, even in one of the 14 ulcer patients (7%), whose lesion was colonized with fungi (Fig
4).
Effects of
Nasogastric
Intubation
Seven patients who had not been intubated previously were examined within 72 hours postburn (Tables 2 and 3); four had the same diffuse superficial mucosal disease of the fundus and body described above. One of these pa-
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Fig
1.—Diffuse
superficial
mucosal disease of fundus identified at
gastroscopy within 72 hours postburn. Erosions and punctate areas of mucosal hemorrhage (dark areas) are present on frugal folds. Linear hemorrhagic lesion (arrow) probably represents trauma from
nasogastric
sump tube.
tients also had antral involvement; two had concomitant erosive disease of the duodenum. All of the patients with mucosal disease had burns involving more than 38% of their total body surface (mean burn size, 59.6%), and le¬ sions were apparent as early as five hour postinjury. Endoscopy was performed in 22 patients after a naso¬ gastric sump tube had been in position for up to 72 hours postburn; four had normal appearing gastric mucosa (Table 3). Although one of these patients had only a 20% total body surface burn, the three others with normal gas¬ tric mucosa had injuries involving more than 67% of their total body surface (mean burn size, 79.8%). All three of these patients had low serum protein levels, and two also had laboratory evidence of disseminated intravascular
Fig 2.—Biopsy specimen of fundic erosion obtained at gastros¬ during second week postburn. Area of glandular regeneration (arrow) is present beneath superficial exúdate of fibrin and neutro¬ phils (hematoxylin-eosin, 100).
copy
Fig 3.—Endoscopie biopsy specimen of gastric lesion obtained within 72 hours postburn showing frank infarction necrosis. Only remnants of glandular epithelium (arrow) are identifiable (hema¬ toxylin-eosin,
100).
coagulation. Eighteen patients had erosive gastritis and an indwell¬ ing nasogastric tube prior to the initial endoscopy (Table 3). While linear mucosal lesions (Fig 1, arrow) frequently suggested a traumatic component of their disease, the dif¬ fuse mucosal congestion, the widespread distribution of discrete lesions, and the frequent, concomitant duodenal disease indicated a more generalized mucosal insult (Fig 5).
Microvascular Thromboses and Disseminated Intravascular Coagulation
Laboratory evidence of disseminated intravascular co¬ agulation was present in five of the 24 patients (21%) eval¬ uated within the initial 72-hour period (Tables 2 and 3). Although three of these patients also had superficial gas¬ tric disease, mucosal biopsy specimens from two of them
failed to demonstrate'microvascular thrombi. Two other patients with laboratory evidence of consumption coagulopathy had normal-appearing stomachs and mucosal
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Fig 4.—Postmortem section of acute gastric ulcer showing necro¬ extending beyond muscularis mucosae. Branching pseudohyphae of Candida are present in superficial necrotic exúdate. Moderate inflammatory response (arrows) has been generated (hematoxylin-eosin, x 100). sis
Fig 5.—Gastroscopy on third day postburn shows nasogastric sump tube (arrow) passing over area of intense superficial mucosal disease. Complete endoscopie examination showed diffuse mucosal disease of fundus, corpus, antrum, and duodenum.
Table 2.—Correlation of Acute Gastric Disease With Intubation, Coagulation Abnormalities, and Serum Protein Levels in 29 Patients Evaluated Within 72 Hours Postburn
No. of Patients
Patient's Status at Initial Endoscopy
Patients With Gastritis
Laboratory Evidence Histologie Evidence of Consumption of Microvascular Thrombi Coagulopathy Yes
No
Yes
Low Total Serum Protein Levels*
No
Low
13
17
Normal
First day postburn With nasogastric tube Without nasogastric tube Second
day postburn
With nasogastric tube Without nasogastric tube Third day postburn With nasogastric tube Without nasogastric tube Total
19
29
Incidence, % *
Less than 6
13 15 4
12 1
22
19 76
21
81
gm/100 ml.
biopsy specimens even though both had burns of greater than 67% of the total body surface, an indwelling naso¬ gastric tube, and total serum protein levels of less than 6 gm/100 ml. Mucosal biopsy specimens from seven other patients with gastric erosions but normal or nondiagnostic clotting studies also showed no microvascular throm¬ boses. Total Serum Protein Levels
Total serum protein levels were above 6.0 gm/100 ml in only four of the 21 tested patients (Table 2); only one pa¬ tient had a serum albumin level above 3 gm/100 ml. Two of the four patients with normal total serum protein levels had burns of greater than 50% of the total body surface
and both had diffuse mucosal disease. The two others had normal-appearing stomachs but had sustained burns of less than 30% of the total body surface. Normal gastric mucosa was present in three other patients with large burns (mean burn size, 63.8% of the total body surface) in spite of low serum protein levels; two of these also had an indwelling nasogastric tube for up to 72 hours before the initial endoscopie examination. As seen in Table 3, patients with normal and diseased gastric mucosa tended to have low serum protein levels. Patients with acute gastric disease, however, had signifi¬ cantly lower protein levels than patients with normal mu¬ cosa (mean, 4.48 gm/100 ml vs 5.72 gm/100 ml; P< .05). Although patients with "gastritis" had larger burns than
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Table 3.—Clinical Characteristics of 29 Patients Evaluated Within 72 Hours Postburn Normal Gastric Mucosa
Acute Gastric Disease No. of Patients Evaluated 22
Clinical Characteristics
Age, yr*
Mean i SD
Range Burn size, % total Mean ± SD
body surfacef
Range Consumption coagulopathy Normal phosphotungstic acid-hematoxylin biopsy specimens
Results
32.5 ± 15.6 18-74
37 ±16.9 16-55
57.6 ± 17.0 38-96
50.1 ±29.1 20-89
4.48 ±1.08 2.4-6.5
5.72 ± .988 4.4-6.8
22
Range Total serum protein, gm/100 Mean ± SD
Results
No. of Patients Evaluated
16
19
& disseminated intravascular
coagulation Intubation prior to endoscopy 22 18 No previous intubation 22 * No significant difference between ages of patients with and without disease ¡sease. t No significant difference between burn sizes of patients with and without di=>c< Difference between patients with and without gastric disease significant at level of .05
patients with intact gastric mucosa (mean, 57.6% vs 50.1% of the total body surface), this difference was not signifi¬ cant, nor was there a significant age difference between both groups of patients. Clinical
Complications
As indicated in Table 1, upper gastrointestinal bleeding originated from gastric lesions in six of the 54 patients (11.1%). In five patients, hemorrhage from diffuse super¬ ficial erosions was shown. This bleeding was easily con¬ trolled with iced saline lavage and was never life-threat¬ ening. Only one patient required vagotomy and subtotal gastrectomy for control of gastric ulcer bleeding. One other patient perforated a gastric ulcer that was not diag¬ nosed premortem. Nineteen patients had multiple gastroscopie examina¬ tions; six were evaluated on three occasions. Eight of these 19 patients showed progression of their mucosal dis¬ ease. The clinical course of each patient had been compli¬ cated by sepsis, respiratory insufficiency, or hypotension. As six other patients convalesced uneventfully, their mucosal disease improved. Complete healing occurred in four patients within three weeks postburn; in two of these, erosive disease recurred after the development of septicemia and hypotension. COMMENT
Retrospective surveys relying on clinically obvious com¬ plications, such as hemorrhage or perforation, identify acute gastric disease in less than 30% of thermally injured patients.7' Gastroduodenoscopy indicates that this disease is much more common in the immediate postburn period than appreciated clinically.17 In this study, superficial ero¬ sive disease of the stomach was present in 83.5% of pa¬ tients with burns involving more than 30% of their total body surface. Acute gastric ulcérations were discovered
later in the
postburn period in 26% (14 of 54 patients) and appeared to represent progression of the early mucosal in¬ jury. Diffuse gastric disease could be asymptomatic even in patients with multiple ulcers. Positive tests for occult blood from nasogastric aspirate or stool were commonly the only clinical signs of the disease process. Disease pro¬ gression followed clinical decompensation of the patient. Although brisk gastric hemorrhage occurred in 11.1% of patients, this bleeding was usually controlled with iced saline lavage. Life-threatening hemorrhage or perfora¬ tion occurred only in patients with an ulcer. The cause of this acute gastric disease is unknown. The frequent, early association of both gastric and duodenal disease (concomitant involvement in 76% of 45 patients with mucosal disease) indicates a common susceptibility to the initial pathogenic insult. The appearance of gastric le¬ sions as early as five hours postburn and the histologie evi¬ dence of infarction necrosis suggest that acute mucosal is¬
chemia, occurring almost immediately postburn, is of etiologic importance. The generalized mucosal pallor and
the results of the morphologic examination of the lesions encountered within 72 hours postburn are similar to the observations made previously in acutely hemorrhaged ani¬ mals.14 Since fibrin thrombi were not evident in any of the mucosal biopsy specimens obtained from patients with erosive disease, including those with laboratory evidence of disseminated intravascular coagulation, microvascular thrombosis did not appear to be an important etiologic factor. The opening of submucosal arteriovenous shunts at the time of thermal injury, however, could have initiated the pathologic process, resulting in focal areas of mucosal injury with associated edema and hemorrhage. The arte¬ riovenous anastomoses in the human stomach are respon¬ sive to control by the autonomie nervous system; sympa¬ thetic stimulation shunts blood away from the mucosa.15"17 In restrained rats, the opening of these shunts produces
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
generalized mucosal ischemia followed by focal tissue damage, possibly related to the local release of vasoactive substances.7 The abrupt shunting of blood away from the gastric mucosa immediately postburn in response to sym¬ pathetic discharge or catecholamine release could crit¬ ically compromise gastric mucosal perfusion in the burned patient. As in the study of Lucas et al,18 the antrum was less fre¬ quently diseased than the fundus and body. A greater he¬ modynamic stability under conditions of "stress,"19-'-'0 a more rapid mucosal regeneration rate,-1 or a reduced energy demand" may make the antrum less susceptible than the more proximal part of the stomach to injury from acute alterations in mucosal perfusion. Serial endoscopie examinations indicated that antral disease improved more rapidly than fundic disease, even in the presence of bile
reflux. The acutely injured gastric mucosa may be predisposed to additional damage by acid, bile salts, or digestive en¬ zymes. Hypoproteinemia and prolonged nasogastric intu¬ bation may augment this damage.1' Total serum protein levels of less than 6 gm/100 ml were present in 81% of tested patients. This hypoproteinemia could contribute to mucosal edema, and a negative nitrogen balance, associ¬ ated with thermal injury, might impair mucosal regenera¬ tion.23 Patients with acute gastric disease tended to have
lower protein levels than patients with nor¬ mal gastric mucosa. The lower protein levels in patients with acute gastropathy might have reflected substantial losses of protein from the injured mucosa. Low serum pro¬ tein levels, however, were not consistently associated with mucosal disease or large burns. Similarly, while it is im¬ possible to determine from this study the entire contribu¬ tion of the nasogastric tube to the mucosal injury, it is evi¬ dent that unintubated patients could develop early diffuse mucosal disease. Although the lesions of tube trauma were frequently identified, the diffuse mucosal congestion with erosions and the concomitant duodenal disease indicated a more generalized mucosal insult. While coagulation abnor¬
significantly
malities, nasogastric intubation, or hypoproteinemia might have enhanced mucosal injury, they were clearly not of primary etiologic importance. Early gastric mucosal injury may well be impossible to prevent after a major burn; however, control of those fac¬ tors that may augment this injury is possible. In this re¬ gard, early vigorous nutritional support and judicious use of nasogastric suction may prove beneficial. Close moni¬ toring of patients with extensive burns, particularly those who develop septic and other complications, will permit identification of disease progression, early diagnosis of perforation or substantial hemorrhage, and timely surgi¬ cal intervention to enhance survival.
References 1. Dupuytren B: Lecons Orales de Clinique Chirurgicale. Paris, Bailliere, 1832, vol 1, p 459. 2. Sevitt S: Duodenal and gastric ulceration after burning. Br J Surg 54:32-41, 1967. 3. Pruitt BA Jr, Foley FD, Moncrief JA: Curling's ulcer. Ann Surg 172:523-538, 1970. 4. Kirksey TD, Moncrief JA, Pruitt BA Jr, et al: Gastrointestinal complications in burns. Am J Surg 116:627-633, 1968. 5. Day SB, MacMillan BG, Altermeier WA: Curling's Ulcer. Springfield, Charles C Thomas Publisher, 1972, p 180. 6. Czaja AJ, McAlhany JC, Pruitt BA Jr: Acute gastroduodenal disease following thermal injury. N Engl J Med 291:925-929,1974. 7. Hase T, Moss BJ: Microvascular changes of gastric mucosa in
the
development of
234, 1973.
stress ulcer in rats.
8. Chiu CJ, McArdle
Gastroenterology
AH, Brown RA,
65:224\x=req-\
et al: Gastric mucosal
changes following burns in rats. Arch Surg 103:147-152, 1971. 9. Guth PH, Kozbur X: Pathogenesis of gastric microcirculatory and mast cell changes in restraint stress. Am J Dig Dis 13:530-535, 1968. 10. Margaretten W, McKay DG: Thrombotic ulcerations of the gastrointestinal tract. Arch Intern Med 127:250-253, 1971. 11. Linder MN, McKay DG: An experimental study of thrombotic ulceration of the gastrointestinal mucosa. Surg Gynecol Obstet
133:21-29, 1971. SR, Wangensteen OH: Role of hemoconcentration
12. Friesen
in production of gastric and duodenal ulcer following tal burns. Proc Soc Exp Biol Med 64:81-85, 1947.
experimen-
13. Carrasquilla C, Weaver A, Amarasinghe DC, et al: Gastroesophageal erosions and ulcerations. Arch Surg 107:447-451, 1973. 14. Harjola PT, Sivula A: Gastric ulceration following experimentally induced hypoxia and hemorrhagic shock. Ann Surg 163:21-28, 1966. 15. Barclay AE, Bentley FH: The vascularization of the human stomach: A preliminary note on the shunting effect of trauma. Br J Radiol 22:62-67, 1949. 16. Barlow TE, Bentley FH, Walder DN: Arteries, veins and arteriovenous anastomoses in the human stomach. Surg Gynecol Obstet 93:657-671, 1951. 17. Womack NA: Blood flow through the stomach and duodenum. Am J Surg 117:771-780, 1969. 18. Lucas CE, Sugawa C, Riddle J, et al: Natural history and surgical dilemma of "stress" gastric bleeding. Arch Surg 102:266\x=req-\ 273, 1971. 19. Richardson RS, Norton LW, Sales JEL, et al: Gastric blood flow in endotoxin-induced stress ulcer. Arch Surg 106:191-195, 1973. 20. Mackie B, Turner MD: Long-term blood flow studies in the gastric submucosa of unanesthetized rats. Arch Surg 103:500-504, 1971. 21. Lawson HH: Gastritis and gastric ulceration. Br J Surg 53:493-496, 1966. 22. Menguy R, Masters YF: Mechanism of stress ulcer. Gastroenterology 66:509-516, 1974. 23. Hunt TK: Injury and repair in acute gastroduodenal ulceration. Am J Surg 125:12-18, 1973.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Injury
MAJ Albert J. Czaja, MC, USA; MAJ Joseph C. McAlhany, Jr., MC, USA; MAJ Willard A. Andes, MC, USA; COL Basil A. Pruitt, Jr., MC, USA
Diffuse erosive "gastritis" was discovered as early as five hours postinjury in 45 of 54 burn patients (83.5%) evaluated by gastroduodenoscopy. Acute ulcers were identified in 14 patients (26%); concomitant duodenal disease was present in 34 patients (76%). Microvascular fibrin thrombi were not demonstrated even though five patients had disseminated intravascular coagulation. Seven patients were examined before nasogastric intubation; four, with a mean burn size of 59.6% total body surface, had diffuse "gastritis." Low total serum protein levels were measured in 81% of tested patients, but were not predictive of mucosal disease. Hemorrhage followed the clinical deterioration of six patients (11.1%); one ulcer perforated. Whereas coagulation abnormalities, nasogastric intubation, and hypoproteinemia may augment mucosal injury, the morphologic and histologic examinations of the lesions suggested a primary ischemic cause resulting from the opening of submucosal shunts or local vasoconstriction.
The development following
of acute erosions of the gastric mu¬ cutaneous thermal injury was first described by Dupuytren in 1832.' Retrospective clinical and postmortem studies2'5 have recognized this erosive "gastritis" as a life-threatening complication in the burn population, occurring with a frequency estimated between 14.4%- and 24.7%."' Endoscopie examinations of the stom¬ ach and duodenum within the immediate postburn period have indicated an even higher incidence of disease." Acute mucosal ischemia secondary to the opening of submucosal cosa
Accepted
a
for publication Dec 15, 1974. From the US Army Institute of Surgical Research, Brooke Army Medical Center, Fort Sam Houston, Tex. Read before the 82nd annual meeting of the Western Surgical Association, San Francisco, Nov 22, 1974. The opinions or assertions contained herein are the private views of the authors and are not to be considered as official or as reflecting the views of the Department of the Army or the Department of Defense. Reprint requests to US Army Institute of Surgical Research, Brooke Army Medical Center, Fort Sam Houston, TX 78234 (Dr. Czaja).
arteriovenous shunts,1-7 8 local vasoconstriction,71 or the de¬ velopment of microvascular thromboses from a dis¬ seminated intravascular coagulation1"" or postburn hemoconcentration1- has been implicated in the cause. An indwelling nasogastric tube (routinely used to decompress the burn patient with ileus), hypoproteinemia, and a nega¬ tive nitrogen balance may augment the mucosal injury.13 To determine the clinical and pathologic characteristics of this acute gastric disease, early and serial endoscopie ex¬ aminations of the stomach and duodenum were performed in thermally injured patients admitted to the US Army Institute of Surgical Research. In 29 patients examined within 72 hours postburn, coagulation studies and mucosal biopsy specimens, specially stained to detect fibrin thrombi, assessed the etiologic importance of microvascular thromboses. The interrelationship between the acute gastric disease, the indwelling nasogastric tube, and hypoproteinemia was also evaluated. METHODS All nonpediatric patients admitted within one week after sus¬ taining a burn of more than 25% of their total body surface were eligible for endoscopy. A history of chronic gastrointestinal symp¬ toms or excessive alcohol or aspirin ingestion eliminated the pa¬ tient from the protocol. Each patient was managed independently by his physician. The informed written consent of the patient or his next of kin was obtained prior to each procedure. Endoscopie examination of the entire stomach and proximal part of the duodenum was usually accomplished within 72 hours postburn, utilizing a fiberoptic panendoscope. This evaluation was repeated during the second and the third week of hospitalization. Other ex¬ aminations were performed whenever indicated clinically. Intra¬ venous administration of diazepam (up to 20 mg), occasionally supplemented by intravenous administration of meperidine hy¬ drochloride (up to 50 mg), was begun immediately before each en¬ doscopie procedure. Photographs documented the mucosal disease.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Table 1.—Correlation of Acute Gastric Disease With Burn Size, Duodenal Disease, and Clinical
% of Total
Body Surface
No. of Patients
Burn 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 90-100 Total
Incidence,
Erosive Gastritis*
Acute Gastric
Ulcerf
Acute Duodenal Disease
Gastric
Hemorrhaget
Complications
Gastric Perforation
Operations for Gastric Disease
0
14 16
12 16
54
45 83.5
/o
* Fundus and body always involved; antrum also diseased t Multiple gastric ulcers in five patients (36%). X Bright red bleeding via nasogastric tube.
14
14 26 in 15
6 11.1
34 63
1 1.86
1 1.86
patients (33%).
Biopsy specimens were obtained from the gastric lesions and the adjacent intact mucosa. These specimens were processed with hematoxylin-eosin and phosphotungstic acid-hematoxylin stains; histologie sections were examined for the presence of microvascular fibrin thrombi. The platelet count, fibrinogen level, pro¬ thrombin time, partial thromboplastin time, and other selected co¬ agulation studies, including fibrin degradation products, serial thrombin time, plasminogen, and factors V and VIII, were deter¬ mined on admission and were reevaluated every two to four days during the first two weeks of hospitalization. These studies were reviewed by a hematologist who formulated a clinical inter¬ pretation of the coagulation changes while unaware of the endo¬ scopie findings. A sudden, marked decrease in both the platelet count and fibrinogen level in conjunction with a prolongation of the prothrombin and partial thromboplastin times suggested the diagnosis of disseminated intravascular coagulation. Serum albu¬ min and total protein levels were monitored during this same interval. Patients with gastrointestinal ileus were routinely decom¬ pressed with a No. 18 F nasogastric sump tube attached to inter¬ mittent low pressure suction. Endoscopy was performed immedi¬ ately before or shortly after nasogastric intubation. Occasionally, the sump tube was in place for up to 72 hours before the initial en¬ doscopie examination. Localized or linear mucosal lesions were at¬ tributed to intubation trauma.
RESULTS
Seventy-nine endoscopie examinations were performed without complication in 54 patients. All patients were male, with three exceptions. Ages ranged from 16 to 76 years (mean age, 33.4 years) and burn sizes ranged from 20% to 96% of the total body surface (mean burn size, 55.6% of the total body surface). Mortality within this study group was 76% (41 of 54 patients). Postmortem ex¬ amination in 36 cases (88% of the patients studied) cor¬ roborated the findings of gastroscopy. Twenty-nine patients were evaluated within 72 hours postburn; one patient was examined as early as five hours postinjury. Twenty-five other patients were studied from four to 14 days after their burn. Mucosal biopsy specimens were obtained from 13 of the patients evaluated within the initial 72-hour period, and coagulation studies were
completed in 24. Serum protein levels were measured dur¬ ing this same period in 21 patients. Acute
Superficial Gastric Mucosal Disease ("Gastritis") and Ulcération
Table 1 indicates the incidence of erosive "gastritis" and acute ulcération, and the relationship of acute gastric disease to burn size, duodenal disease, and clinical compli¬ cations. Mucosal lesions were not discovered in patients with burns involving less than 30% of their body surface. As indicated in Table 2, acute gastric mucosal disease was present in 76% of patients evaluated within 72 hours postinjury. Erythematous petechial lesions, confluent macular areas of mucosal hemorrhage, and punctate erosions on a pale mucosa were identified within the fundus and body of the stomach (Fig 1). The antrum was also involved in 15 of the 45 patients with gastric disease (33%). Concomitant duodenal disease was present in 34 of the patients with "gastritis" (76%). Acute gastric ulcers were not discovered before 72 hours postburn and always occurred in areas of severe superficial mucosal disease. Five of the gastric ul¬ cer patients (36%) had multiple craters. The histologie examination of the macular lesions showed microvascular congestion, edema, and focal muco¬ sal hemorrhage. Superficial epithelial necrosis above the muscularis mucosae, usually with an exúdate of fibrin and neutrophils, characterized the erosions. Mucosal regenera¬ tive activity was occasionally evident (Fig 2). Frank mucosal infarction was also encountered (Fig 3). Mucosal necrosis extending beyond the muscularis mucosae distin¬ guished the ulcers. Inflammatory reaction around the ul¬ cer was only mild to moderate, even in one of the 14 ulcer patients (7%), whose lesion was colonized with fungi (Fig
4).
Effects of
Nasogastric
Intubation
Seven patients who had not been intubated previously were examined within 72 hours postburn (Tables 2 and 3); four had the same diffuse superficial mucosal disease of the fundus and body described above. One of these pa-
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Fig
1.—Diffuse
superficial
mucosal disease of fundus identified at
gastroscopy within 72 hours postburn. Erosions and punctate areas of mucosal hemorrhage (dark areas) are present on frugal folds. Linear hemorrhagic lesion (arrow) probably represents trauma from
nasogastric
sump tube.
tients also had antral involvement; two had concomitant erosive disease of the duodenum. All of the patients with mucosal disease had burns involving more than 38% of their total body surface (mean burn size, 59.6%), and le¬ sions were apparent as early as five hour postinjury. Endoscopy was performed in 22 patients after a naso¬ gastric sump tube had been in position for up to 72 hours postburn; four had normal appearing gastric mucosa (Table 3). Although one of these patients had only a 20% total body surface burn, the three others with normal gas¬ tric mucosa had injuries involving more than 67% of their total body surface (mean burn size, 79.8%). All three of these patients had low serum protein levels, and two also had laboratory evidence of disseminated intravascular
Fig 2.—Biopsy specimen of fundic erosion obtained at gastros¬ during second week postburn. Area of glandular regeneration (arrow) is present beneath superficial exúdate of fibrin and neutro¬ phils (hematoxylin-eosin, 100).
copy
Fig 3.—Endoscopie biopsy specimen of gastric lesion obtained within 72 hours postburn showing frank infarction necrosis. Only remnants of glandular epithelium (arrow) are identifiable (hema¬ toxylin-eosin,
100).
coagulation. Eighteen patients had erosive gastritis and an indwell¬ ing nasogastric tube prior to the initial endoscopy (Table 3). While linear mucosal lesions (Fig 1, arrow) frequently suggested a traumatic component of their disease, the dif¬ fuse mucosal congestion, the widespread distribution of discrete lesions, and the frequent, concomitant duodenal disease indicated a more generalized mucosal insult (Fig 5).
Microvascular Thromboses and Disseminated Intravascular Coagulation
Laboratory evidence of disseminated intravascular co¬ agulation was present in five of the 24 patients (21%) eval¬ uated within the initial 72-hour period (Tables 2 and 3). Although three of these patients also had superficial gas¬ tric disease, mucosal biopsy specimens from two of them
failed to demonstrate'microvascular thrombi. Two other patients with laboratory evidence of consumption coagulopathy had normal-appearing stomachs and mucosal
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Fig 4.—Postmortem section of acute gastric ulcer showing necro¬ extending beyond muscularis mucosae. Branching pseudohyphae of Candida are present in superficial necrotic exúdate. Moderate inflammatory response (arrows) has been generated (hematoxylin-eosin, x 100). sis
Fig 5.—Gastroscopy on third day postburn shows nasogastric sump tube (arrow) passing over area of intense superficial mucosal disease. Complete endoscopie examination showed diffuse mucosal disease of fundus, corpus, antrum, and duodenum.
Table 2.—Correlation of Acute Gastric Disease With Intubation, Coagulation Abnormalities, and Serum Protein Levels in 29 Patients Evaluated Within 72 Hours Postburn
No. of Patients
Patient's Status at Initial Endoscopy
Patients With Gastritis
Laboratory Evidence Histologie Evidence of Consumption of Microvascular Thrombi Coagulopathy Yes
No
Yes
Low Total Serum Protein Levels*
No
Low
13
17
Normal
First day postburn With nasogastric tube Without nasogastric tube Second
day postburn
With nasogastric tube Without nasogastric tube Third day postburn With nasogastric tube Without nasogastric tube Total
19
29
Incidence, % *
Less than 6
13 15 4
12 1
22
19 76
21
81
gm/100 ml.
biopsy specimens even though both had burns of greater than 67% of the total body surface, an indwelling naso¬ gastric tube, and total serum protein levels of less than 6 gm/100 ml. Mucosal biopsy specimens from seven other patients with gastric erosions but normal or nondiagnostic clotting studies also showed no microvascular throm¬ boses. Total Serum Protein Levels
Total serum protein levels were above 6.0 gm/100 ml in only four of the 21 tested patients (Table 2); only one pa¬ tient had a serum albumin level above 3 gm/100 ml. Two of the four patients with normal total serum protein levels had burns of greater than 50% of the total body surface
and both had diffuse mucosal disease. The two others had normal-appearing stomachs but had sustained burns of less than 30% of the total body surface. Normal gastric mucosa was present in three other patients with large burns (mean burn size, 63.8% of the total body surface) in spite of low serum protein levels; two of these also had an indwelling nasogastric tube for up to 72 hours before the initial endoscopie examination. As seen in Table 3, patients with normal and diseased gastric mucosa tended to have low serum protein levels. Patients with acute gastric disease, however, had signifi¬ cantly lower protein levels than patients with normal mu¬ cosa (mean, 4.48 gm/100 ml vs 5.72 gm/100 ml; P< .05). Although patients with "gastritis" had larger burns than
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
Table 3.—Clinical Characteristics of 29 Patients Evaluated Within 72 Hours Postburn Normal Gastric Mucosa
Acute Gastric Disease No. of Patients Evaluated 22
Clinical Characteristics
Age, yr*
Mean i SD
Range Burn size, % total Mean ± SD
body surfacef
Range Consumption coagulopathy Normal phosphotungstic acid-hematoxylin biopsy specimens
Results
32.5 ± 15.6 18-74
37 ±16.9 16-55
57.6 ± 17.0 38-96
50.1 ±29.1 20-89
4.48 ±1.08 2.4-6.5
5.72 ± .988 4.4-6.8
22
Range Total serum protein, gm/100 Mean ± SD
Results
No. of Patients Evaluated
16
19
& disseminated intravascular
coagulation Intubation prior to endoscopy 22 18 No previous intubation 22 * No significant difference between ages of patients with and without disease ¡sease. t No significant difference between burn sizes of patients with and without di=>c< Difference between patients with and without gastric disease significant at level of .05
patients with intact gastric mucosa (mean, 57.6% vs 50.1% of the total body surface), this difference was not signifi¬ cant, nor was there a significant age difference between both groups of patients. Clinical
Complications
As indicated in Table 1, upper gastrointestinal bleeding originated from gastric lesions in six of the 54 patients (11.1%). In five patients, hemorrhage from diffuse super¬ ficial erosions was shown. This bleeding was easily con¬ trolled with iced saline lavage and was never life-threat¬ ening. Only one patient required vagotomy and subtotal gastrectomy for control of gastric ulcer bleeding. One other patient perforated a gastric ulcer that was not diag¬ nosed premortem. Nineteen patients had multiple gastroscopie examina¬ tions; six were evaluated on three occasions. Eight of these 19 patients showed progression of their mucosal dis¬ ease. The clinical course of each patient had been compli¬ cated by sepsis, respiratory insufficiency, or hypotension. As six other patients convalesced uneventfully, their mucosal disease improved. Complete healing occurred in four patients within three weeks postburn; in two of these, erosive disease recurred after the development of septicemia and hypotension. COMMENT
Retrospective surveys relying on clinically obvious com¬ plications, such as hemorrhage or perforation, identify acute gastric disease in less than 30% of thermally injured patients.7' Gastroduodenoscopy indicates that this disease is much more common in the immediate postburn period than appreciated clinically.17 In this study, superficial ero¬ sive disease of the stomach was present in 83.5% of pa¬ tients with burns involving more than 30% of their total body surface. Acute gastric ulcérations were discovered
later in the
postburn period in 26% (14 of 54 patients) and appeared to represent progression of the early mucosal in¬ jury. Diffuse gastric disease could be asymptomatic even in patients with multiple ulcers. Positive tests for occult blood from nasogastric aspirate or stool were commonly the only clinical signs of the disease process. Disease pro¬ gression followed clinical decompensation of the patient. Although brisk gastric hemorrhage occurred in 11.1% of patients, this bleeding was usually controlled with iced saline lavage. Life-threatening hemorrhage or perfora¬ tion occurred only in patients with an ulcer. The cause of this acute gastric disease is unknown. The frequent, early association of both gastric and duodenal disease (concomitant involvement in 76% of 45 patients with mucosal disease) indicates a common susceptibility to the initial pathogenic insult. The appearance of gastric le¬ sions as early as five hours postburn and the histologie evi¬ dence of infarction necrosis suggest that acute mucosal is¬
chemia, occurring almost immediately postburn, is of etiologic importance. The generalized mucosal pallor and
the results of the morphologic examination of the lesions encountered within 72 hours postburn are similar to the observations made previously in acutely hemorrhaged ani¬ mals.14 Since fibrin thrombi were not evident in any of the mucosal biopsy specimens obtained from patients with erosive disease, including those with laboratory evidence of disseminated intravascular coagulation, microvascular thrombosis did not appear to be an important etiologic factor. The opening of submucosal arteriovenous shunts at the time of thermal injury, however, could have initiated the pathologic process, resulting in focal areas of mucosal injury with associated edema and hemorrhage. The arte¬ riovenous anastomoses in the human stomach are respon¬ sive to control by the autonomie nervous system; sympa¬ thetic stimulation shunts blood away from the mucosa.15"17 In restrained rats, the opening of these shunts produces
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
generalized mucosal ischemia followed by focal tissue damage, possibly related to the local release of vasoactive substances.7 The abrupt shunting of blood away from the gastric mucosa immediately postburn in response to sym¬ pathetic discharge or catecholamine release could crit¬ ically compromise gastric mucosal perfusion in the burned patient. As in the study of Lucas et al,18 the antrum was less fre¬ quently diseased than the fundus and body. A greater he¬ modynamic stability under conditions of "stress,"19-'-'0 a more rapid mucosal regeneration rate,-1 or a reduced energy demand" may make the antrum less susceptible than the more proximal part of the stomach to injury from acute alterations in mucosal perfusion. Serial endoscopie examinations indicated that antral disease improved more rapidly than fundic disease, even in the presence of bile
reflux. The acutely injured gastric mucosa may be predisposed to additional damage by acid, bile salts, or digestive en¬ zymes. Hypoproteinemia and prolonged nasogastric intu¬ bation may augment this damage.1' Total serum protein levels of less than 6 gm/100 ml were present in 81% of tested patients. This hypoproteinemia could contribute to mucosal edema, and a negative nitrogen balance, associ¬ ated with thermal injury, might impair mucosal regenera¬ tion.23 Patients with acute gastric disease tended to have
lower protein levels than patients with nor¬ mal gastric mucosa. The lower protein levels in patients with acute gastropathy might have reflected substantial losses of protein from the injured mucosa. Low serum pro¬ tein levels, however, were not consistently associated with mucosal disease or large burns. Similarly, while it is im¬ possible to determine from this study the entire contribu¬ tion of the nasogastric tube to the mucosal injury, it is evi¬ dent that unintubated patients could develop early diffuse mucosal disease. Although the lesions of tube trauma were frequently identified, the diffuse mucosal congestion with erosions and the concomitant duodenal disease indicated a more generalized mucosal insult. While coagulation abnor¬
significantly
malities, nasogastric intubation, or hypoproteinemia might have enhanced mucosal injury, they were clearly not of primary etiologic importance. Early gastric mucosal injury may well be impossible to prevent after a major burn; however, control of those fac¬ tors that may augment this injury is possible. In this re¬ gard, early vigorous nutritional support and judicious use of nasogastric suction may prove beneficial. Close moni¬ toring of patients with extensive burns, particularly those who develop septic and other complications, will permit identification of disease progression, early diagnosis of perforation or substantial hemorrhage, and timely surgi¬ cal intervention to enhance survival.
References 1. Dupuytren B: Lecons Orales de Clinique Chirurgicale. Paris, Bailliere, 1832, vol 1, p 459. 2. Sevitt S: Duodenal and gastric ulceration after burning. Br J Surg 54:32-41, 1967. 3. Pruitt BA Jr, Foley FD, Moncrief JA: Curling's ulcer. Ann Surg 172:523-538, 1970. 4. Kirksey TD, Moncrief JA, Pruitt BA Jr, et al: Gastrointestinal complications in burns. Am J Surg 116:627-633, 1968. 5. Day SB, MacMillan BG, Altermeier WA: Curling's Ulcer. Springfield, Charles C Thomas Publisher, 1972, p 180. 6. Czaja AJ, McAlhany JC, Pruitt BA Jr: Acute gastroduodenal disease following thermal injury. N Engl J Med 291:925-929,1974. 7. Hase T, Moss BJ: Microvascular changes of gastric mucosa in
the
development of
234, 1973.
stress ulcer in rats.
8. Chiu CJ, McArdle
Gastroenterology
AH, Brown RA,
65:224\x=req-\
et al: Gastric mucosal
changes following burns in rats. Arch Surg 103:147-152, 1971. 9. Guth PH, Kozbur X: Pathogenesis of gastric microcirculatory and mast cell changes in restraint stress. Am J Dig Dis 13:530-535, 1968. 10. Margaretten W, McKay DG: Thrombotic ulcerations of the gastrointestinal tract. Arch Intern Med 127:250-253, 1971. 11. Linder MN, McKay DG: An experimental study of thrombotic ulceration of the gastrointestinal mucosa. Surg Gynecol Obstet
133:21-29, 1971. SR, Wangensteen OH: Role of hemoconcentration
12. Friesen
in production of gastric and duodenal ulcer following tal burns. Proc Soc Exp Biol Med 64:81-85, 1947.
experimen-
13. Carrasquilla C, Weaver A, Amarasinghe DC, et al: Gastroesophageal erosions and ulcerations. Arch Surg 107:447-451, 1973. 14. Harjola PT, Sivula A: Gastric ulceration following experimentally induced hypoxia and hemorrhagic shock. Ann Surg 163:21-28, 1966. 15. Barclay AE, Bentley FH: The vascularization of the human stomach: A preliminary note on the shunting effect of trauma. Br J Radiol 22:62-67, 1949. 16. Barlow TE, Bentley FH, Walder DN: Arteries, veins and arteriovenous anastomoses in the human stomach. Surg Gynecol Obstet 93:657-671, 1951. 17. Womack NA: Blood flow through the stomach and duodenum. Am J Surg 117:771-780, 1969. 18. Lucas CE, Sugawa C, Riddle J, et al: Natural history and surgical dilemma of "stress" gastric bleeding. Arch Surg 102:266\x=req-\ 273, 1971. 19. Richardson RS, Norton LW, Sales JEL, et al: Gastric blood flow in endotoxin-induced stress ulcer. Arch Surg 106:191-195, 1973. 20. Mackie B, Turner MD: Long-term blood flow studies in the gastric submucosa of unanesthetized rats. Arch Surg 103:500-504, 1971. 21. Lawson HH: Gastritis and gastric ulceration. Br J Surg 53:493-496, 1966. 22. Menguy R, Masters YF: Mechanism of stress ulcer. Gastroenterology 66:509-516, 1974. 23. Hunt TK: Injury and repair in acute gastroduodenal ulceration. Am J Surg 125:12-18, 1973.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - San Diego User on 06/09/2015
