Annals of the MBC - vol. 4 - n' 4 - December 1991

SUMMARY. An empiric antibiotic treatment was made with th i rd-gene ration cephalosporins (ceptazidime-aminoglycoside), in 18 critical burns patients with clinical sepsis without bacteriological confirmation; the evaluation of plasmatic levels and M.I.C. (minimum inhibitory concentration) of ceptazidime was made. The successes and failures of the this dual antibiotic treatment, the high incidence of sepsis by Staphylococcus epidermidis in relation to Pseudomonas aerugmosa (20%), and the selection of resistances to aminoglycosides and beta-lactam antibiotics were discussed. We recommend treatment with aminoglycoside-vancomycin antibiotic in the case of existence of sepsis without bacteriological confirmation.

The critical burns patient is one of the most serious in medical -surgical pathology, as he is a prototype of the immunosuppressed patient (Gelfand 1984), since equally the burn in itself, the numerous surgical operations, and the blood and plasma transfusions lead to a transitory secon arimmunodeficiency syndrome. There are there ore frequent septic complications which constitute, according to Curreri et al. (1980), the main cause of mortality, even though the local and systemic treatments are correct. The principal microorganisms which provoke these sepses are Staphyloccocus and Pseudomonas (MacMillan et al., 1986), and thirdly Enterobacteriaceae. For this reason an empiric treatment with aminoglycoside and third-generation cephalosporins was started, due to the good results obtained, according to several publications (Blint et al., 1983, Abbas et al., 1983, Wittman et al., 1981).

A) All the critical burns patients with a TBB surface > 40% and/or of electrical origin, with risk factors such as post-inhalation respiratory insufficiency, renal failure, existence of cardiopathy or pulmonary obstructive illness, and age over 65 or under 15 years, and who for these reasons were considered "high risk" on admission, were included. The number of patients with clinical sepsis (with or without later bacteriological confirmation) was 18, in the first 8 months of 1986. They were all admitted to the Critical Burns Unit of "La Paz" Hospital. The empiric treatment consisted of 2 gm of ceptazidime i.v. every 6 hours and an aminoglycoside (generally tobramycin) in conventional doses. The prospective study and the evaluation of the treatment were made considering a "success" to be the non-existence of micro-organisms in the blood on completion of treatment; "superinfectlon" the finding of new germs different from the original ones; and "failure" the persistence of the same micro-organisms responsible for the sepsis.

B) M.I.C. (minimum inhibitory concentration) of ceptazidime against isolated micro-organisms. 64 micro-organisms isolated from critical care patients in 1985 were used: 18 Pseudomonas aeruginosa, 10 Staphylococcus aureus, 12 Escherichia coli or Proteus sp. and 24 S.E.K. group (Serratia, Enterobacter, Klebsiella). The M.I.C. was obtained by means of the introduction into Mueller-Hinton agar plates of ceptazedime in increasing concentrations (ratio: 2), from 0.5 to 128 microgr/ml The previous microorganisms, at a concentration of 10/ml, were inoculated in the plates by means of the Steers applicator, and the smallest concentration of anti-microbial agent which inhibited the bacterial growth in 48 hours was considered to be the M.I.C., as carried out in a previous work (Herruzo et al., 1985). In the same way two micro-organisms, Staphylococcus aurcus ATTCC 6538 and Pseudomonas aeruginosa ACC 15442, were used as a double check.

C) Evaluation of ceptazidime levels in patients' plasma. This was done by microbiological method (Edberg and Sabath 1980) with a strain of Enterobacter cloacae resistant to aminoglycoside and susceptible to ceptazidime. Blood was extracted to determine these le vels half an hour before and after each of 6 doses of the antibiotic combination. 9 cases were analysed. The pattern concentrations and their zones of inhibition were introduced into a Mackintosh Plus computer for processing with the Cricket programme, which calculates the equation of regression between the variables and displays them graphically, allowing us to calculate the problem levels either by substituting the corresponding inhibition zone in the equation or by extrapolation of the zone from the graph.

The characteristics of the 18 patients as well as their infections and evaluation of appropriate treatments are displayed in Table 1. The following facts are clear:

Average age: 33.9 +/- 3.5 years
Average TBB: 48.7 +/- 4.8%
Kidney failure: 2 cases
Average duration of empiric treatment: 6.7 +/-
Sepsis post-treatment: 11 cases (6 1 % of total)
Post-treatment sepsis was due to:

• - S. epidermidis: 18 cases (73% of total patients)
• - P. aeruginosa: 2 cases (18% of total patients)
• - Candida sp.: 1 case (9.1% of total patients)

Negative blood culture post-treatment: 6 cases (33%)
"Success" in empiric treatment: 6 cases (33% of total)
"Superinfection" in empiric treatment: 2 cases (11% of total)
"Failure" in empiric treatment: 2 cases (11% of total)
"Failure" or supermfection" in empiric treatment: 7 cases (39% of total)
"Unknown":1 case (immediate decease)
Secondary treatment: 10 cases (91% of total post-treatment sepsis)
Secondary treatment with vancomycin: 7 cases (70% of secondary treatments)
Success of secondary treatment with vancomycin: 5 cases (71% of total vancomycin treatments)
Superinfection in the secondary treatment with vancomyein: 2 cases (29% of total vancomycin treatments)

NOTE: The micro-organisms included in the "previous culture` concept were not known at the time of instigating the therapy because it is an empiric treatment. The M.I.C. of the 64 micro-organisms and its double-checking for ceptazidime are laid out in Table 11. M.I.C. 50 and M.I.C. 90 are distinguished by~dotted lines ( ... ) and ( --- ). It is noticeable that M.I.C. 50 is very low (<1 microgr/ml) in Staphyloccocus, E. coli and Proleus, low (<4 microgr/ml) in Pseudomonas, and moderate (<16 microgr/ml) in the K.E.S. group.

M.I.C. 90 is high (-~128 and 64 microgr/ml respectively) in Staphylococcus and Pseudomonas, moderate (<16 microgr/ml) in the K.E.S. group, low (<8 microgr/ml) in E. coli and Proteus. Fig. I represents the curve-pattem and its corresponding equation in order to calculate the ceptazidime levels in burn patient plasma. The shaded area represents where the 9 cases are to be found: peak levels less than 4 microgr /ml in 8 of the 9 patients and 10 microgr/ml in the 9th. The "valley" levels were all less than 2 microgr/ml, undetectable by the microbiological method.

What stands out in this study is the very low mortality in spite of the prediction of mortality on admission to the Unit. This is due to the better treatment, both surgical and general, of critical burns patients, compared to that given when the probit indexes were established in the Unit (197 1). Of all the cases which constitute a success in empiric treatment, one was selected, P. aeruginosa (non poly-R), in the digestive tract as a consequence of this treatment. Successes were obtained in Enterobacter sp. or S. epidermidis sepsis (3 cases).
The failure, number 9, was however, effective in eradicating the enterobacteria (Proteus), although of course the accompanying micro-organism Candida was not affected. In the last 2 successful cases the micro-organism responsible for the sepsis was unknown. What is noticeable in the treatment failures is the high frequency of S. epidermidis in the sepsis, which is probably due to an intragranulocytic defect in the burns patients, which makes them incompetent for effective phagocytosis of the micro-organism (Loose and Turinsky, 1979), together with the existence of entrances (catheters i.v.), which must be kept open for a long time for the patients' sustenance. These entrances are essential for the blood invasion of the micro-organisms, in general scarcely pathogenous in burns, so it is improbable that they originate from the septic burn. The P. aeruginosa responsible for these post-treatment sepses (20%) can obtain access to the blood stream by infection of the burn or of the area around the catheter. Although the initial M.I.C. in the majority of the micro-organisms isolated from the different patients in the Unit is moderate or low, the empiric treatment used created resistant mutations which have produced infections, especially Pselidomonas aeruginosa (poly-R): 2 sepses (Cases 4 and 18) and a failure (Case 9). There is a total resistance selection in these strains to all aminoglycosides and beta-lactam antibiotics in general, as has also been obtained by other authors using third-generation cephalosporins (Sanders and Sander, 1985). Perhaps the reason for this resistance is the small concentration in the blood of antibiotics used in burned patients, as has been shown for ceptazldime, and as has been described by other authors in aminoglycosides (Zaske et at., 1976; Zaske et al., 1978., Loisat et al., 1978), However the levels reached with vancomycin are generally useful for controlling staphylococcal sepsis (due to reduced M.I.C. to this antibiotic), in spite of the high renal clearance of the antibiotic, as described by Brater (Brater et al., 1986).
Antibiotic treatment with aminoglycosidevancomycin is therefore recommended in suspected sepsis without bacteriological confirmation, the dosage being adapted to the pharmacokmetic clearance of these products. In the same way, in confirmed bacteriological sepsis, originated by sensitive Pseudomonas and enterobacteria, ceptazidime could be used if the pharmacokinetic is determined for adaptation of the dosage.