Annals of Burns and Fire Disasters - vol. Xl - n. 1 - March 1998

SELECTIVE GASTROINTESTINAL DECONTAMINATION AND BURN WOUND SEPSIS

Shalaby H.A.,¹ Higazi M.,¹ El Far N.²

¹Plastic and Reconstructive Surgery Unit' and Microbiology Department ²
Tanta Faculty of Medicine, Tanta University, Tanta, Egypt

SUMMARY. Many experimental studies have assessed the role of gastrointestinal decontamination, among other factors, on bacterial and endotoxin translocation in burn patients. In this study, 256 burned patients were allocated randomly into three groups. Patients in group 1, in whom there was no gastrointestinal decontamination, were used as control; group 11 patients were subjected to treatment with colistine sulphate, co-trimoxazole, and nystatin; and group III patients received selective gastrointestinal decontamination (SGD) plus xanthine oxidase inhibitors. Systemic antibiotics were given to any patient if indicated by clinical or laboratory signs only. Early feeding with a fibre-rich diet was given to all patients. The results indicated that SGD markedly reduced the chance of bactaeraerma and wound colonization, especially by gram-negative enteric rods. The addition of xanthine oxidase inhibitors to SGI) had no significant effect on the results.

The past two decades have witnessed the emergence of a new syndrome, termed multiple organ failure, which today represents the number one cause of death in surgical intensive care units in severely burned patients.' The attendant morbidity and mortality is anything from 30 to 100%, depending on the number of organ systems involved! In most cases, a septic focus - such as the wound, the genitourinary tract, or the respiratory tract - is readily identifiable. In other cases, the source of sepsis is not apparent! During the past decade, clinical and laboratory investigations have pointed towards the gastrointestinal tract as a reservoir of organisms which in certain circumstances leave the confines of the gastrointestinal tract lumen to produce a septic state.
The first investigation linking the gastrointestinal tract to the lethality of traumatic shock was performed by Jacob Fine in 19652 In a shock model, he showed the gut to be the source of both endotoxins and invasive gramnegative bacteria. In 1979, Berg and Garlington' coined the term "translocation", which they defined as "the passage of viable bacteria through the epithelial mucosa into the lamina propria and then to the mesenteric lymph nodes, and possibly other organs".
The balance of the microbial flora indigenous to the gastrointestinal tract is critical in determining which species of bacteria translocate. In rodents whose guts were decontaminated and then colonized with a single E. coli species, translocation increased. Restoring a normal flora reduced the translocation! In the process of treating infections of the primary burn injury, alterations in the gastrointestinal microtiora will occur, even with use of parenteral as opposed to oral antibiotics.'Endotoxin has been found to cause damage to mouse mucosa: when administered by an intraperitoneal route, it was associated with increased bacterial translocation in burned animals.
The composition of the diet may alter the flora and also affect the integrity of the mucosal lining. The combination of protein malnutrition and endotoxin resulted in a higher incidence of bacterial translocation in mice." Alverdy et al. 12 studied rats fed on a normal enteral diet with oral total parenteral nutrition (TPN) solution or by intravenous TPN solution. The animals fed orally on a normal diet showed a lower incidence of translocation. Alverdy et al." found that diets rich in fibres and glutarnine were associated with the least degree of bacterial translocation in rats.
Thermal injury is often associated with shock and decreased visceral blood flow. In an experimental model, haemoffhagic shock was associated with bacteria] translocation." Allopurinol, a xanthine oxidase inhibitor, reduced bacterial translocation from 61 to 10%. It appears that shock-induced bacteria] translocation from the gut may be mediated by free-oxygen radicals, generated by activation of the xanthine oxidase system.
Other factors predisposing to translocation are immunosuppression, 16 abdominal irradiation," obstructive jaundice," and intestinal obstruction.
A common criticism of the concept of bacteria] translocation is that it is a phenomenon produced in the laboratory under artificial conditions." The present study was designed to evaluate the effect of selective gastrointestinal decontamination (SGD), with or without allopurinol, on the incidence of burn wound sepsis.

This prospective study included 256 patients, after exclusion of patients who died within 48 h post-burn (resuscitation period). The patients presented a burn surface area of 25% or more in adults and 15% or more in children.
The clinical evaluation of each patient included the causative agent, associated chronic illness, the presence of hypovolaemic shock at admission and its duration, inhalation injury, and the size and depth of the burn.
Resuscitation was effected with intravenous lactated Ringer's solution 4 ml/kg/%TBSA. Fluids were considered satisfactory if urine output was 0.5-1 ml/kg/h. Enteral feeding initiated early, as soon as the gut functioned. If the patient could not tolerate oral feeding, tube feeding was used. The diet used was rich in fibres (bran-rich bread, whole-com bread, whole beans, and vegetables rich in cellulose fibre).
Local wound care was performed with a bath and cleanser solution, local povidine-iodine 10%, and silver sulphadiazine 1% cream. The trunk, head and neck were dressed open twice daily, while the limbs were closed, with dressings changed once daily.
Early wound excision and autografting were performed within 3-5 days in the hands, face and feet. Delayed grafting, after 2-3 weeks, was the normal routine in other body areas.
Systemic antibiotics were given if there were clinical or laboratory signs of sepsis. Bacteriological surveillance of the burn wound and blood cultures were performed twice weekly, starting from the day of admission. The wound was considered to be colonized by micro-organisms when the same micro-organism was isolated from two consecutive samples, while the susceptibility of potential pathogenic micro-organisms was determined by the agar diffusion test, using susceptibility test tables (Neosensitabs, Roscodiagnostica, Toastrup, Denmark).
The patients were allocated randomly to three groups: group 1 (85 patients) - control group treated as specified above; group 11 (90 patients) - SG1). Patients were treated like the control group, plus; a) colistine sulphate - 150,000 unit/kg/day; b) co-trimoxazole tablets of 80 mg trimethoprime and 400 mg sulpharnethoxazole - 2 tablets twice daily (syrup for children); c) nystatin - adult dose 500.00 IU 4 times daily - child's dose 100.00 IU 4 times daily; group 111 (81 patients) - SG1), plus xanthine oxidase inhibitor (allopurinol) 100 mg tablets three times daily, child's dose 50 mg three times daily.

Results

Two hundred and seventy-nine patients were admitted to our Burn Unit with burns of 25% or more BSA in adults and of 15% or more in children. Twenty-three patients died within the first 48 h, i.e. the resuscitation phase, mostly from hypovolaernic shock: these were excluded from the study. The remaining patients were enrolled into one of three groups, using the admission number for randomization.
The statistical analysis of the data regarding these patients on admission revealed no significant statistical differences. These data included the age of the patient, the incidence of inhalation injury, the percentage and depth of the burn, associated chronic diseases, and the incidence of admission shock (systolic blood pressure below 90 mm Hg. Sex and the cause of the burn (flame or hot liquids) were not found to present any difference with regard to control and the other two groups.
The first positive blood culture was recorded on postburn day 8.22 ± 5.68 days in the control group, compared with 13.67 ± 8.35 days in the SG1) group and 13.83 ± 8.67 days in the SG1) plus allopurinol group. There was a significant statistical difference between control and the other two groups (p < 0.001). There was no significant statistical difference between the two treatment groups (groups II and III).
When admission systolic blood pressure was above 90 ram Hg, gram~negative enteric rods were not detected in the blood culture during the first week in any of the three groups. However, in patients admitted with hypovolaemic shock, the incidence of gram-negative bacteraemia was 48.3 % in the first group, 13.3 % in the second group, and 14.3% in the third group. There was a significant statistical difference between group 1 and each of the groups II and III.
During the second week, the incidence of bacteraemia was 47.8% (41 patients) in the first group, of whom nine (22%) were gram-negative enteric rods. Bacteraemia was 10% in the second group with two cases (20%) of gramnegative enteric rods. In the third group, bacteraemia was 7.4% (six patients), of whom two (33.3%) were gramnegative enteric rods. The difference in the incidence of bacteraemia was significant between group 1 and the other two groups. Also, the incidence of gram-negative enteric rods was significant between control and treatment groups.
During the third and fourth weeks, the incidence of bacteraemia and gram-negative enteric rods decreased in the three groups, but also with a still significant difference between control and the treatment groups.
Invasive burn wound infection manifested itself in the three groups respectively on mean post-burn day 8.07 + 5.22, 14.88 ± 7.60 and 14.28 ± 6.70. There was a highly significant statistical difference between group 1 and each of the other two (p < 0.001).
During the first week, the main organisms detected in the burn wound were gram-positive cocci. However, gramnegative enteric rods were found in seventeen patients in the first group, eight in the second, and seven in the third group. The difference was statistically significant between group 1 and the other two groups.
During the second week, the total incidence of burn wound infection declined. However, gram-negative enteric rods were still high in group 1. Also, fungal wound infection was detected in thirteen wounds in group 1, four in group 11, and two in group Ill, also with statistically significant differences.
In the third and fourth weeks, the incidence of detection of gram-positive cocci increased in the three groups while that of gram-negative enteric rods decreased. The difference between group 1 and each of groups 11 and 111 was however still significant. Fungal wound infection was detected in only three cases in group 1, and no case was detected in the other two groups.
The incidence of mortality in burn patients admitted was 17.8% (50 out of 279). After exclusion of patients who died in the resuscitation phase (23 patients), the mortality rate was found to be 10.6% (27 out of 256). In the first group it was 15.3% (13 out of 85), compared with 7.8% in the second group and 8.6% in the third. The difference between group 1 and each of the other two groups was statistically significant.
The mean number of days of hospitalization in the first group was 30.11 ± 9.70. In group 11 it was 23.86 ± 9.42 days, and in group three 24.99 ± 9.64 days. The statistical difference was high between groups 1 and 11 and also between groups 1 and 111 (p < 0.001). No such difference was found between groups II and III.

The relationship between the microflora of the digestive tract and that of the burn wound has been reported repeatedly, but only a few investigators have attempted to influence burn wound colonization by suppression of gastrointestinal microflora.
Translocation of bacteria and endotoxins after major burns has been well demonstrated in experimental models.'' Selective digestive tract decontamination is a wellestablished method for the prevention of infection in granulocytopaenic patients." It has also been used in other imnumocompromised patients, particularly those in intensive care units '22 patients subjected to liver transplants 2' and patients undergoing prolonged mechanical ventilation.
The regimen used for selective gut decontamination in this study achieved a delay in wound colonization and an overall reduction in the incidence of infections, due especially to enteric gram-negative bacilli and fungal infection, compared with the control group.

The first positive blood culture in this study showed a highly significant difference between that of the control group and those of both the other two groups. Also, the first day of positive wound culture and infection showed the same difference. The difference between groups 11 and 111 was statistically insignificant. The addition of allopurinol did not therefore increase the improvement achieved by gut decontamination alone. In a retrospective study, Manson et al.` studied the SG1) regimen in burn patients, reporting that the triple SG1) regimen (polymyxin B, cotrimoxazole and amphotericin B) resulted in better eradication of aerobic gram-negative flora from the gut compared with a group treated with polymyxin B alone.
Candidaemia and Candida wound infection was found to be more common in the control group than in the other two groups treated in this series. The rate of Candida wound and blood stream infections has been recorded to be on the increase in recent years?' This is due to an increase in the number of critical cases treated in intensive units and of immunosuppressed patients.
The results of our study suggest that the delay in burn wound infection with selective gut decontamination permitted earlier wound closure and reduced the incidence of severe burn wound sepsis and mortality. The mortality rate in the control group was 15.3%, compared with 7.8% and 8.6% in the second and third groups. The difference between the control and the other two groups was statistically significant. The mortality rates in reported series vary considerably, from a low of 4.4% in Saudi Arabia` to a high of 22% in Zaria in Nigeria. 29 In Alexandria,` Egypt, it was found to be 21.2%.
The improvement in survival can be explained by a better understanding of burn pathophysiology, improved laboratory facilities, infection control, nutritional support, and early oral feeding. The addition of selective gut decontamination has added to the improvement in the care of burn patients and limited the use of systemic antibiotics to cases diagnosed by clinical or laboratory signs as having infection of their wounds. The reduction in hospital stay, the limited use of systemic antibiotics, and early enteral feeding have contributed to the reduction of financial costs and to the patients' sense of well-being.

RESUME. Beaucoup de rechercheurs ont évalué le rôle de la décontamination gastro-intestinale, parmi d'autres facteurs, sur la translocation des bactéries et des endotoxines dans les patients brûlés. Les Auteurs de cette étude ont considéré 256 patients brûlés divisés au hasard en trois groupes. Le premier groupe, employé comme témoin, ne contenait pas de cas de décontamination gastro-intestinale, le deuxième était composé de patients traités avec le sulfate de colistine, le co-trimoxazole et la nystatine; et le troisième etait composé de patients traités moyennant la décontamination gastro-intestinale sélective (DGS) plus des inhibiteurs de xanthine oxydase. Des antibiotiques systémiques ont été administrés au patient dans les cas où ils étaient indiqués seulement par des indicateurs cliniques ou de
laboratoire. L'alimentation précoce, avec une diète riche de fibres, a été employée chez tous les patients. Les résultats ont indiqué que la DSG a réduit notamment la fréquence de la bactériémie et la colonisation de la lésion, en particulier par les bâtonnets entériques à Gram négatif. L'addition de l'inhibiteur de la xanthine oxidase à la DGS n'a pas produit aucun effet signifiant sur les résultats.

BIBLIOGRAPHY

1. Epstein M.D., Banducci D.R., Manders E.K: The role of the gastrointestinal tract in the development of burn sepsis. Plast. Reconstr. Surg., 90: 524, 1992.
2. Garrico C.J., Meakins J.L., Marshal J.C. et al.: Multiple organ failure syndrome. Arch. Surg., 121: 196, 1986.
3. Deitch E.A., Berg R.D.: Endotoxin, but not malnutrition, promotes bacterial translocation of the gut flora in burned mice. J. Trauma,27: 171, 1987.
4. Epstein M.D., Tchervenkov J.L, Alexander JW., Inoue S., Kuroiwa K., Jacobs P.: Effect of intraluminal antibiotics on translocation of Candida albicans in burned guinea pigs. Burns, 16: 105, 1990.
5. Fine J.: Current status of the problem of traumatic shock. Surg. Gynecol. Obstet., 120: 537, 1965.
6. Berg R.D., Garlington A.W.: Translocation of certain indigenous bacteria from the gastrointestinal tract to the mesenteric lymphnodes and other organs in a gnotobiotic mouse model. Infect.Immun., 23: 403, 1979.
7. Berg R.D.: Promotion of the translocation of enteric bacteria from the gastrointestinal tracts of mice by oral treatment with penicillin,clindamycin or metronidazole. Infect. Immun., 33: 854, 1981.
8. Wells C.L., Maddaus M.D., Reynold C.M., Jechorek R.P., Simmons R.L.: Role of anaerobic flora in the translocation of aerobic and facultatively anaerobic intestinal bacteria. Infect. Immun., 55: 2689,1987.
9. Deitch E.A., Winterton J., Borg R.D.: Effect of starvation, malnutrition and trauma on the gastrointestinal tract flora and bacterial translocation. Arch. Surg., 122: 1019, 1987.
10. Deitch E.A., Ma L., Ma W.J., Grisham M.B., Gragner D.N., Specian R.D., Berg R.D.: Inhibition of endotoxin-induced bacterial translocation in mice. J. Clin. Invest., 84: 36, 1989.
11. Deitch E.A., Bridges W.: Effect of stress and trauma on bacterial translocation from the gut. J. Surg. Res., 42: 536, 1987.
12. Alverdy J.C., Aoys E., Moss G.S.: Total parenteral nutrition promotes bacterial translocation from the gut. Surgery, 104: 185, 1988.
13. Alverdy J.C., Aoys E., Moss G.S.: Effect of commercially available chemically defined liquid diets on the intestinal microfiora and bacterial translocation from the gut. J. Parent. Ent. Nutt., 14: 1, 1990.
14. Baker J.W., Deitch E.A., Li M., Berg R.D., Specian R.D.:Hemorrhagic shock induces bacterial translocation from the gut. J. Trauma, 28: 896, 1988.
15. Deitch E.A., Bridges W., Baker J., Ma JW., Ma L., Grisham M.B.,Granger D.N., Specian R.D., Berg R.: Hemorrhagic shock induced bacterial translocation is reduced by xanthine oxidase inhibitor or inactivation. Surgery, 104: 191, 1988.
16. Berg R.D., Wommack E., Deitch E.A.: Immunosuppression and intestinal bacterial overgrowth synergistically promote bacterial translocation. Arch. Surg., 123: 1359, 1988.
17. Guzman-Stein G., Bonsack M., Liberty J., Delaney J.P.: Abdominal radiation causes bacterial translocation. J. Surg. Res., 46: 104, 1989.
18. Deitch E.A., Sitting K., Li M., Berg R.D., Specian R.D.: Obstructive jaundice promotes bacterial translocation from the gut. Am. J. Surg., 159: 79, 1990.
19. Deitch E.A.: Bacterial translocation: is it of clinical significance? Gastroenterology, 98: 243, 1990.
20. Epstein M.D., Tchervenkov J.I., Alexander J.W.: Increased gut permeability following burn trauma. Arch. Surg., 126: 198, 1991.
21. De Vries H.G., Sleijfer D.T., Mulder N.H.: Bacteriological aspects of selective intestinal decontamination of the digestive tract as a method of infection prevention in granulocytopenic patients. Antimicrob. Chemother., 19: 813, 1981.
22. Stoutenbeck C.P., Van Saene H.K., Miranda DR.: The effect of oropharyngeal decontamination using topical nonabsorbable antibiotics on the incidence of nosocomial respiratory tract infection in multiple trauma patients. J. Trauma, 27: 357, 1987.
23. Wiesner R.H., Kram R.A., Rakala J.: Selective bowel decontamination to prevent gram-negative and fungal colonization following liver transplantation. Hepatology, 6: 1207, 1986.
24. Claesner H.A., Volaord E.J., Van Saene H.F.: Long-term prophylaxis of infection by selective intestinal decontamination in leucopenia and in mechanical ventilation. Rev. Infect. Dis., 9: 295, 1987.
25. Van Saene J.M., Van Saene H.F., Stoutenbeck C.P.: Influence of feces on the activity of antimicrobial agents used for decontamination of the alimentary canal. Scand. J. Infect. Dis., 17: 295, 1985.
26. Manson W.L., Sauer E.W.: Selective intestinal decontamination of the digestive tract. A tool for infection prophylaxis in burns. Ann. Medit. Burns Club, 7: 88, 1994.
27. Wey S.B., Mori M., Pfaller M.A.: Risk factors for hospital acquired candidemia. A matched case-control study. Arch. Intem. Med., 149: 2349, 1989.
28. Jamal U.S., Ardawi M.S., Asly A.R. et al.: Pediatric injuries in the Jeddah area of Saudi Arabia. A study of 197 patients. Burns, 16: 34, 1990.
29. Mabogunje O.A., Lawrie J.H.: Burns in adults in Zaria, Nigeria.Burns, 14: 308, 1988.
30. Mostafa M.F., Borhan A., Abdulla A.F. et al.: A retrospective study of 5505 burned patients admitted to Alexandria Burns Unit. Ann. Medit. Burns Club, 3: 269, 1990.