The mortality data are collected in Tab. 4, which shows that the two intermediate years differ significantly form the first (but not from the fourth) in non-immediate mortality. There were no significant differences in age and burned body surface area percentage among the fatal cases over the four years. In 1987 the number of patients with electrical burns was greater than in the preceding years, although the ditTerence is only significant for the comparison with 1984, due to the small size of the sample. Tab. 5 summarizes the manipulations performed on the patients. There was an increase in time for most manipulations, which was greatest for assisted ventilation and blader catheterization. Surgical interventions also increased 100% from 1984 1986-1987. Tab. 6 shows the incidence, aetiology and site of infections. These data are graphically summarized in Figs. 1, 2 and 3. A striking observation is the decline of infection, both when considered as cumulative incidence and as incidence density, during the last two years, and even more so in 1987, when the infections per patient per month represent only one quarter of the figure for 1984. Furthermore, the most frequent infection in this last year (in spite of the decline as compared to the previous ones) is that in "other soft tissues" (with non-significant differences between 1984-87), due to the even more drastic reduction of infections of other sites (particularly bum P < 0.001). As to the aetiology of infections, there was a very marked fall of all organisms in burn infections (p < 0.001), and only Pseudomonas retain an appreciable incidence (although this is only 25% of that during the first two years). Enterobacteria, the main cause of burn infections, disappeared in 1987. There has also been an inversion of the relative frequencies of both types of staphylococcl (S. aureus and S. epidermidis), as S. aureus, which was more frequent in previous years, disappeared, and only S. epidermidis had a small representation in the last year of the study.
The cases of sepsis indicate an overall reduction to one sixth of the initial incidence (p < 0.001); the main causative agent is now Pseudomonas for even though its incidence is lower than in the first two years it has become predominant due to the greater reduction (p < 0.001) of the other organisms (except anaerobes and fungi: non-significant difference).
In 1984, sepsis due to Gram-positive cocci was caused mainly by S. aureus. Over the following two years S. epidermidis predominated, in spite of the overall reduction of this type of infection, and in 1987 both types of staphylococci were isolated with similar low frequency.
Infections of the soft tissues were those which benefited least from the therapeutic measures introduced in 1987, with a non-significant reduction, although S. aureus was no longer represented as an aetiological agent and there was a slight increase in the incidence of Gram-negative rods.
Infection of the urinary tract also showed a reduction (p < 0.001 between 1984-87), while other infections hardly need to be mentioned; foremost among these are the infections of the respiratory tract.
When calculating the overall germ-specific incidence of infection, regardless of site, there is a clear decline for all micro-organisms in 1987, as compared to previous years: 330-1340% for enterobacteria, 10-1230% for anaerobes, 1240-5300% for S. aureus, 10-370% for fungi, 0-240% for Pseudomonas, 30-430% for S. epidermidis, and 300-470% for other organisms (mainly S. faecalis).
Consumption of antibiotics, which is represented in Figs. 4, 5 and 6, declines in parallel to the incidence of infection with the exception of years 1985-86, where consumption was high due to increased use of third-generation cephalosporins and aminoglycosides.
Finally, Tab. 7 compares several indices of the four years, standardized with 1987 as the reference year. This comparison shows a "saving" of 46-114 infections and of 9-28 million pesetas (admissions plus antimicrobial drugs) as an effect of the introduction of the new therapeutic measures.

The foremost and most remarkable finding of this retrospective review is the dramatic decrease in the number of infections achieved during the last two years of the study period, when double prophylaxis (intestinal and perioperatory) was applied, and most particularly in 1987, when perioperatory prophylaxis was reduced to a valid minimum (two doses) and intestinal prophylaxis was carried out with three drugs and associated to local action on the pharynx.
These procedures achieved a striking reduction of all infecting micro-organisms, and the reduction is rather less striking only in the case of infections of other tissues due to Pseudomonas, in spite of the particular attention paid to the protective measures (27).
The great impact that prophylaxis achieved on the incidence of infections demonstrates the endogenous origin of the immense majority of these complications in this type of patient (28); only in a few cases of nonfermenting bacilli was an endogenous infection suspected, although we were unable to demonstrate the mechanism of transmission (hands of nursing staff, inert carriers, etc.).
We also had some failures in intestinal decolonization, in spite of the established prophylaxis, which were caused in most cases by multiresistant

Fig. 1 - Distribution of Hospital Infection (measured in incidence density). Fig. 2 - Aetiology of Hospital Infection (incidence in density of incidence). Fig. 3 - Aetiology of Hospital Infection (continuation) (incidence i density of incidence). Fig. 4 - Density of infection incidence in relation to antibiotic use (expressed in DDD per 100 Bed-days).

This may be explained by the fact that the dose of colimycin used for prophylaxis was too low for this organism, although it was fully effective for controlling enterobacteria. For this reason, in 1988 we introduced prophylaxis with colimycin at a dose of 600-800 ing/day (18), using oral norfloxacin or tobramycin if bacteria were selected which are naturally resistant to colimycin, such as Proteus or Serratia.
As shown in Fig. 3, the incidence of fungal infections increased (non-significantly) in 1986. This was a consequence of the increased consumption of antibiotics, mainly third-generation cephalosporins. The incidence of fungi as aetiological agents again diminished in 1987, for two reasons: reduction of the use of selected antibiotics (for instance, two instead of five doses for perioperatory prophylaxis) and association of an antifungal agent (Nistatin) for decolonization of the intestinal reservoir, although the dose we used was lower than that suggested by Stovtenbeek (17) and Clasener (18); even so, it was highly effective in preventing infection by these organisms. In 1988 the dose of Nistatin was increased when Candida was isolated from stools, and so far all detected colonizations have been eradicated.
Staphylococcus epidermidis infections are mainly associated to intravenous fluids and catheters, and increase in parallel with these manipulations. The greater attention materials has led infections and aureus has develonment paid to these manoeuvres and to a reduction in the number of to a change in their aetiology, where S. replaced S. epidermidis, but the of such infection must be further reduced, an aim we hope to achieve through the new measures (local application of 0.5% chlorhexidine cream onto the area of insertion of catheter: the initial results of this procedure have been optimal).

Fig. 5 - Antibiotic use. Description. Fig. 6 - Betalactarn. Antibiotic use.

In spite of the large proportion of patients requiring assisted ventilation (33% in 1987), the incidence of respiratory infections was minimal; we believe this finding to decolonization with demonstrated in a study intensive care unit (19).
The striking reduction of the infections cannot be explained solely by the greater recourse to surgery. The number of surgical interventions increased over the first three years of the study period, but in the fourth year, with a similar proportion of surgical interventions per patient, there was a reduction of the infections to half the number of the previous year. Furthermore, the decline of infections achieved between 1985 and 1986 was much greater than the increase in the frequency of surgery, which was similar to that from 1984 to 1985. These observations underline the importance of global decolonization in the early management of the bum patient.
The increased mortality in 1987 as compared to the two preceding years, although non-significant, might be due to the greater age of the patients who died (2-9 years), to the worse state on admission (increased frequency of aspiration syndrome) and/or to multisystem failure, as suggested by the increase be related to pharyngeal hexitidine, as already carried out in a different (3.4 to 8.2 times) of the proportion of patients requiring assisted ventilation, and to the greater number of patients with electrical bums (4-14).
With regard to antimicrobial drug consumption, in 1987 there was a reduction in both the overall consumption and the consumption per patient; there was also a change in antimicrobial usage, from the predominantly therapeutic usage during the first years of the study to the mainly prophylactic usage in 1987.
The increased expenditure on antimicrobials observed in 1985-86 was primarily due to the use of first- and third-generation cephalosporins (Figs. 5 and 6) which, in the case of perioperatory prophylaxis (first-generation cephalosporins), might imply a link-up with that of a second intervention, if this took place three or four days after the previous one.
This problem was obviated in 1987, when only two doses were given (before and after surgery), so that even though a greater number of perioperatory prophylaxis regimens were given, both the overall usage of antimicrobials and the selection of resistant strains in the intestine decreased.
Finally, when translating the balance of the infection rates in the unit into "pesetas", a considerable saving effect becomes evident (9 to 29 million pesetas, according to the various years), together with a further saving that is not easily amenable to economical measurements, i.e. that of suffering, of sequelae and also of mortality, which would have been the consequence of the 46 to 101 infections that did not occur.
We therefore stress the great interest of the global prophylactic regime used (Tab. 2), together with the two modifications described in the text, for reducing, together with surgery, the number of infections of all origins in patients sustaining bum injuries.

RÉSUMÉ. Les Auteurs ont étudié la fréquence de la mortalité et de l'infection hospitalière (et en même temps la consommation des antibiotiques et l'épargne économique) chez une Unité de Brûlures, pendant une période de 4 ans, par rapport aux diverses modifications introduites dans la gestion thérapeutique. La mortalité était réduite après le deuxième an, et la fréquence de l'infection était plus basse tous les ans (au quatrième an elle était 25% par rapport au premier an). Il y a eu aussi des modifications pour ce qui concerne l'étiologie et le site de l'infection: l'infection locale causée par la brûlure a diminué à 25% dans la période, et seulement P. aeruginosa et S. epidermis ont maintenu une fréquence sensible; les infections généralisées ont eu une évolution semblable à celle des infections locales des brûlures; l'infection aux voies urinaires a eu une fréquence très basse pendant toute la période; les autres infections du tissu mou sont devenues prédominantes au quatrième an à cause d'une réduction des infections plus classiques. L'évolution de la consommation des antibiotiques n'est pas corrélée au taux d'infection, à cause de l'emploi très large des céphalosporines de la troisième génération pendant les années centrales de la période; une comparaison, après la standardisation directe, entre les évènements pendant la période complessive et ceux de l'an final indique une épargne économique considérable (9-28 millions de pesetas), comme conséquence de l'introduction des nouvelles techniques thérapeutiques.

BIBLIOGRAPHY

1. McManus W.F., Goodwins CW., Manson A.D., Pruitt B.A.: Bum wound infection. Journal of Trauma: 753-756, 1981.
2. Yurt R.W., Shires G.T.: Bums. In: Mandell G.L., Douglas R.G. and Bennett J.E., Eds., "Principles and practices of infectious diseases", Second ed., 628-631, New York, 1985.
3. Lutterman A., Dacso C.C., Curreri P.W.: Infections in bum patients. American Journal of Medicine, 81 (suppl.): 45-52, 1986.
4. Gelfand J.A.: Infections in bum patients: A paradigm for cutaneous infection in the patient at risk. American Journal of Medicine, 76 (Suppl.): 158-163, 1989.
5. Alexander C.L., Ogle C.K., Stinnett J.D., McMillan B.G.: A sequential prospective analysis of immunologic abnormalities and infection following severe thermal injury. Annals of Surgery: 809-816, 1978.
6. Bjornson A.B., Alterneier W.A., Bjornson H.S.: Complement, opsonin and the immune response to bacterial infection in burned patients. Annals of Surgery, 191: 323-329, 1980.
7. McCabe W.P, Rebuck J.W., Kelly A.P., Ditmars D.M.: Leukocyte response as a monitor of immunodepression in bum patients. Archives of Surgery, 106: 155-159, 1973.
8. Grogan J.B.: Suppressed in vitro chernotaxis of burn neutrophil. Journal of Trauma, 16: 985-988, 1976.
9. Warden J.D., Mason A.D., Pruitt B.A.: Evaluation of leukocyte chernotaxis in vitro in thermally injury patients. Journal of Clinical Investigation, 54: 1001-1004, 1974.
10. McMillan B.G., Holder F.A., Alexander J.V.: Infections of bum wounds. In: Bennett J.V. and Brachman P.S., Eds., oHospital Infections>>, 465-482, Seconded., Little Brown, Boston, 1986.
11. Herruzo Cabrera R., Garcia Caballero J., Lenguas F., Pintado R., Garcia Torres V., Rey Calero J.: Estudio epidemiol6gico de la infecei6n hospitalaria en un centro de quemados durante los ahos 1984-85: Comunicaci6n XI International Congress of Preventive and Social Medicine, Madrid, 1986.
12. Demling R.H.: Improved survival after massive bums. Journal of Trauma, 23: 179-184, 1983.
13. Engnav L.H., Heimbach D.M., Reus J.L., Haman T.J., Marvin S.A.: Early excision and grafting versus nonoperative treatment of bum of indeterminate depth: a randomized prospective study. Journal of Trauma, 23: 1001-1004, 1983.
14. Herruzo Cabrera R., Garcia Torres V., Rey Calero J.: Estudio de la actividad bacteriostàtica y bactericida de 7 cremas antisépticas y evaluaciôn de su poder de penetraciôn mediante un modelo in vitro. Comurlicaciàn al Il Congresso SEIMC, Mallorca, 1986.
15. Herruzo Cabrera R., Pintado Otero R., Rey Calero J., Garcia Torres V.: Complicaciones infecciosas en enfermos quemados. Revista Iberoamericana de Cirugia Pldstica, 14: 257-264, 1988.
16. Van der Vaaij D.: The colonization resistance of the digestive tract in experimental animals and its consequences for infection prevention, acquisition of new bacteria and the prevention of spread of bacteria between cage mates. in: Van der Vaaij D., Verhoef J., Eds., "New criteria for antimicrobial therapy: maintenance of tract colonization resistance", 43-53, Excerpta M6dica, Amsterdam, 1979.
17. Stovtenbeek C.P., Van Sacne H.K.F., Miranda D.R., Zandstra D.F.: The effect of selective decontamination of the digestive tract on colonization and infection rate in multiple trauma patients. Intensive Care Medicine, 10: 185-192, 1989.
18. Clasener H.A.L., Vollaard, Van Saene H.K.F.: Longterm prophylaxis of infection by selective decontamination in leukopenia and in mechanical ventilation. Review of Infectious Diseases, 2: 295-238, 1987.
19. Rodriguez Artalejo F., Herruzo Cabrera R., Garcia Caballero J., Rey Calero J.: Methodologic approaches for the study of pneumonias in intensive care units. XI Congress of Preventive and Social Medicine, Madrid, 1986.
20. Herruzo Cabrera R., Garcia Torres V., Rey Calerc, J.: Control de la Infecci6n en Quemados. Revista Iberoamerjcan~ Cirugia Pldstica, 14: 189-196, 1988.
21. Rothman K.J.: in "Epidemiologia Modema", Diaz de Santos, Ed., 29-41, Madrid, 1987.
22. Kennedy S.L., Forbes M.B., Bann C.: Antibiotic use in U i S. hospitals in 198 1. American Journal of Hospital Pharmacy, 40: 797-801, 1983.
23. Carraco J.L.: "El m6todo estadistico en la investigaci6n m6dica", 167-174, Karpos, Madrid, 1986.
24. Lilienfeld A.M., Lilienfeld D.E.: "Fundamentos de Epidemiologia", 66-68, Fondo Educativo Interamericano, Mexico, 1983.
25. Bennett J.V., Brachman P.S., Eds.: "Hospital Infections", 335-354, Little Brown, Boston, 1979.
26. Halley R.W.: "Managing hospital infection control for cost-effectiveness", American Hospital Publishing, AHA, 1986.
27. Palmer M.B.: "Infection Control. A policy and procedure manual", W.B. Saunders, London, 1984.
28. Bennett J.V., Brachman P.S., Eds.: "Hospital infection", 465-487, Little Brown, Boston, 1986.