F. Gouin% L. Papazian", C. Martin*, J. Albanese*, O. Durbec*, Y. Domarf, P. Veyssier% J. Leroy'', J. J. Gres' and C. Rollin'

Journal of Antimicrobial Chemotherapy (993) 32, Suppl. B, 25-24 A non-comparative study of the efficacy and tolerance of cefepime in combination with amikacin in the treatment of severe infections in patients in intensive care F. Gouin% L. Papazian", C. Martin*, J. Albanese*, O. Durbec*, Y. Domarf, P. Veyssier% J. Leroy'', J. J. Gres' and C. Rollin' "Hopital Sainte Marguerite, Service Anesthesie-Reanimation, 27 bid Sainte-Marguerite, 3274 Marseille Cedex 9; b H6pital Nord, Marseille; C C. H.G. Compiegne; d Hdpital C. Nicolle, Rouen; "Bristol-Myers Squibb, France Patients in intensive care units (ICUs) are at increased risk of developing nosocomial infections. This is of special concern in the immunocompromised patient, particularly with regard to multiresistant pathogens. We evaluated the effectiveness of cefepime 2 g bd in combination with amikacin 7-5 mg/kg bd for the treatment of severe bacterial infection in 8 ICU patients, including 3 patients with nosocomial lower respiratory tract infections (LRTI) (mean age, 5 years). Ninety-six percent (8/3) of the LRTI patients required respiratory assistance and 2% (4/3) had associated septicaemia/bacteraemia. Eighty-four per cent (95/3) had clinical signs of sepsis and 35% (39/3) had features of septic shock. The mean Simplified Acute Physiologic Score (SAPS) was 2 at inclusion. Seventy-nine patients with LRTI were clinically and bacteriologically evaluable. The causative pathogens were representative of those usually isolated in ICUs: Staphylococcus aureus (9%); Pseudomonas aeruginosa (4%); and Klebsiella, Enterobacter and Serratia spp. (7%). The clinical cure rate was 86% (68/79) while the pathogen eradication rate was 9% (7/7). Of the patients with associated septicaemia/ bacteraemia, 89% (8/9) of the pathogens were eliminated. Cefepime-amikacin combination therapy was well tolerated; two patients discontinued treatment due to rashes. Combination therapy with cefepime 2 g bd and amikacin 75 mg/kg bd appears safe and effective for the treatment of nosocomial pneumonia in patients hospitalized in ICUs. Further comparative controlled studies are justified. Introduction Cefepime is a new cephalosporin, the spectrum of activity of which extends to both Gram-positive and Gram-negative pathogens. It is active against streptococci (excepting enterococci), methicillin-susceptible staphylococci and most strains of Enterobacteriaceae and Pseudomonas spp. (Kessler el al., 985). This is related to its structure and physicochemical properties which enable it to penetrate the outer membrane more rapidly, together with a greater affinity for penicillin-binding proteins. It induces less resistance to /Mactamases than cefotaxime and ceftazidime, both in vitro (Phelps et al., 986; Fung-Tome et al., 988) and in vivo (Marchou et al., 987). Compared with ceftazidime, cefepime has been shown to be more active in various experimental models of respiratory tract infections caused by Streptococcus pneumoniae and Staphylococcus aureus, while its efficacy against Klebsiella pneumoniae and 35-7453-93/32B25+ S8./ 25 '(., 993 The British Society for Antimicrobial Chemotherapy

26 F. Gouin et al. Pseudomonas aeruginosa is equivalent to that of ceftazidime (Kessler et al., 985; Tomatsu et al., 986; Masuyoshi et al., 989). Cefepime is weakly bound (6%) to proteins and is excreted primarily (82%) in the urine. Its serum half-life is between 2 and 2-5 h and its serum concentrations between and 2 h after iv or im administration of a dose of 2 g are above the MICs of the most commonly encountered pathogenic bacteria. Thus the pharmacokinetic data suggest that administration every 2 h is justified (Barbhaiya et al., 987, 99). Moreover, cefepime and amikacin can be co-administered without dosage adjustment since it has been demonstrated that there is no pharmacokinetic interaction between these two antibiotics (Barbhaiya et al., 992). The purpose of this study was to evaluate the efficacy and tolerance of cefepime 2 g bd in combination with amikacin in the treatment of severe infections in patients hospitalized in intensive care units (ICUs). Patients Patients and methods The protocol for this multicentre study permitted the inclusion of patients 8 years of age or older with clinical signs and symptoms of severe infection requiring combined antibiotic therapy. Infections were either documented before empirical treatment or subsequently. The exclusion criteria were as follows: allergy to cephalosporins, penicillins or aminoglycosides; haemodialysis or peritoneal dialysis; pregnancy or lactation; the necessity for other concomitant antibiotics; leucopenia (granulocytes < -5 x 9 /L); previous enrolment in the study; severe infection of the central nervous system; HIV disease with pneumocystosis or herpes simplex infection; documented or suspected endocarditis; anaerobic infections; cystic fibrosis or pneumonia distal to a bronchogenic carcinoma; severe burns ( > 2%) with respiratory complications; and any type of prosthesis if this was infected. Study design The study design was non-comparative and multicentre; it was conducted in four French ICUs over a 2-year period. The protocol was approved by the Ethics Committee of each centre. Each patient (or next of kin) gave informed consent before inclusion in the study. Treatment Cefepime was administered iv or im at a dosage of 2 g every 2 h in combination with amikacin 7-5 mg/kg every 2 h for a maximum of 28 days. Amikacin assays were carried out regularly and the dosages adjusted particularly for patients with renal insufficiency. Follow-up The patients were monitored daily. Bacteriological specimens were obtained on entry into the trial and susceptibility to the two study drugs was determined by the disc diffusion method. Further specimens were obtained during and at the end of antibiotic

Cefepime and amikacin in ICU patients 27 treatment. Blood samples were also collected for haematological and biochemical safety monitoring. Definition of responses All patients were evaluable for safety analysis. Therapeutic efficacy was assessed after a minimum of 4 days of treatment and was based on the clinical and bacteriological evaluations performed before, during and after completing treatment. Pneumonia was evaluable when the following criteria were met: (i) sepsis syndrome; (ii) purulent bronchial secretions; (iii) alveolar shadowing on chest X-ray; (iv) bacteriological documentation obtained by blind bronchial sampling (Rouby et al., 992), protected specimen brush (Fagon et al., 989) or bronchoalveolar lavage (Papazian et al., 993). The clinical response was classified as satisfactory (resolution of all features of infection at the end of treatment) or unsatisfactory (persistence, worsening or recurrence of one or more symptom or signs of infection). The bacteriological response was classified in terms of eradication (no pathogens isolated after treatment), persistence (isolation of the initial pathogen after treatment) or new infection (presence of a new pathogen during or after treatment or the presence of the same pathogen at another site). Results One hundred and eighteen patients were included in the study; 3 patients had nosocomial respiratory tract infections. The other five cases included one urinary tract infection, one biliary tract infection, one septicaemia of unknown origin, one postoperative wound infection and one infected retroperitoneal haematoma associated with a urinary tract infection. These five cases were cured by the cefepime-amikacin regimen and are not included in the efficacy analysis of the 3 cases of nosocomial respiratory tract infection. Evaluability Of the 3 patients with nosocomial respiratory tract infections, 79 (7%) were evaluable for efficacy from both a clinical and a bacteriological standpoint. The reasons for excluding the other 34 patients from the efficacy analysis are summarized in Table I. The most frequent causes of non-evaluability were the lack of bacteriological documentation (n = 7) and the isolation of a resistant pathogen (n = 8). Demographics Of the 3 patients included, 86 (76%) were male and 27 (24%) were female. The distribution of the 79 evaluable patients, 59 male (75%) and 2 female (25%) was similar. The mean age was 5 years (range 8-85) for all 3 patients and 5 years (range 8-79) for the 79 evaluable patients. Treatment All 3 patients were treated with cefepime for an average of -3 days (range -24) and with amikacin for 8-9 days (range -2). The 79 evaluable patients were treated

28 F. Gouin et al. Table I. Reasons for non-evaluability No. of non-evaluable patients % total patients Reason (n = 34) studied Reasons related to choice of therapy Resistant pathogen 7 (5-9) Concomitant antibiotic therapy (-8) Premature discontinuation of amikacin (-8) Reasons not related to choice of therapy No pathogens isolated 5 (3) Expectorated sputum specimens (-8) Erroneous exclusion 2 (-7) Duration of treatment ^ 4 days 2 (-7) Doubtful bacteriological documentation (-8) Concomitant vancomycin for Gram-positive cocci although sensitive to cefepime 3 (2-5) Concomitant antibiotic (8) with cefepime for an average of -8 days (range 4-24) and with amikacin for -2 days (range 2-2). Microbiology One hundred and fifty-five pathogens isolated before treatment were tested for susceptibility to cefepime and amikacin. Of the isolated strains, 95% (47/55) were susceptible or of intermediate susceptibility to cefepime. Eighty-two per cent of the isolated strains (27/55) were also susceptible or of intermediate susceptibility to amikacin. The 28 pathogens resistant to amikacin included 27 Gram-positive isolates (25 streptococci and two staphylococci) and one Haemophilus influenzae (/Mactamase-positive). Complicating factors All patients were managed in ICUs. The mean Simplified Acute Physiologic Score (SAPS) was 2 on trial entry for the 3 patients and for the 79 evaluable patients; 76% of the patients had a SAPS of or more (Table II). A total of 8 patients (96%) required respiratory assistance and 33 patients (29%) were 65 years of age or older. Moreover, 4 patients (2%) were septicaemic. The major underlying diseases in all patients were cancer (7), cardiovascular disease (29) and chronic respiratory disease (6). Sepsis and septic shock Ninety-five (84%) of the 3 patients and 66 (84%) of the 79 evaluable patients met the criteria for the sepsis syndrome as defined by the American College of Chest Physicians/Society of Critical Care Medicine (992), i.e. the presence of at least two of the following four signs: (i) fever ( > 38 C) or hypothermia ( < 36 C); (ii) tachypnoea ( > 2 breaths/min); (iii) tachycardia (> 9 beats/min); (iv) leucocyte count > 2xlO 9 /L or <4xlO 9 /L. If the definition of tachypnoea is not restricted to

Cefepime and amikacin in ICU patients 29 Table II. Simplified acute physiologic scores (SAPS) No. (%) of patients SAPS all patients evaluable patients 4-9 -4 5-2 2-25 Total 27 (24) 56 (49) 27 (24) 3 (3) 2 (25) 37 (47) 22 (28) 3() 79 () > 2 breaths/min and is applied to all of the ventilated patients, the number of patients meeting the criteria for the sepsis syndrome then becomes (97%) of the 3 patients and 76 (96%) of the 79 evaluable patients. If the definition of tachypnoea is restricted to > 2 breaths/min, 39 (35%) of the 3 patients and 24 (3%) of the 79 evaluable patients were also hypotensive (systolic blood-pressure < 9 mm Hg and/or needed inotropic support), thereby meeting the criteria for septic shock. These rates were 38% (43/3) and 32% (25/79), respectively, if the broad definition of tachypnoea is used i.e. if all patients ventilated are considered to meet the criteria of tachypnoea. The incidence of each of these signs is given in Table III. Bacteriological documentation An initial bacteriological specimen was obtained for each patient included in the study. Of the 3 respiratory cases, 98 (87%) had positive culture results. In nine of the total cases, the same pathogen was also isolated from a blood culture. In the evaluable patients, samples were obtained by deep blind bronchial aspiration (33), protected specimen brush (9) or bronchoalveolar lavage (27). Clinical and bacteriological responses Of the 3 patients, 79 were evaluated for efficacy. The clinical and bacteriological findings are given in Tables IV and V. At the end of treatment, the SAPS in the evaluable patients was reduced from an average of 2 to 6. Eleven clinical failures were Table III. Presence of signs of sepsis syndrome and shock Sign No. (%) patients all patients evaluable patients Fever ( > 38 C) or hypothermia ( < 36 C) Tachypnoea ( > 2 breaths/min) Tachypnoea or ventilation Tachycardia ( > 9 bpm) Leucocyte count > 2 x 9 /L or < 4 x IO 9 /L Hypotension ( < 9 mm Hg) or use of inotropes Sepsis syndrome Refractory shock 96 (85) 23 (2) 6 (94) 9 (8) 7 (62) 44(39) 95 (84) 39 (35) 67 (85) 6 (2) 72 (9) 6 (76) 49 (62) 26 (33) 66 (84) 24 (3)

2 F. Gouin et al. Table IV. Clinical response rates No. (%) of patients Diagnosis satisfactory unsatisfactory Acute pneumonia 65(86) (4) Acute exacerbation of chronic obstructive pulmonary disease 3 Total 68(86) II (4) observed; six of these patients had septic shock at trial entry. These were nosocomial pneumonias caused by S. aureus (five), P. aeruginosa (three) and Enterobacter spp. (three). Of the nine patients for whom the study treatment was altered to another antibiotic or combination, modified treatment was ineffective in six cases. Twenty-two deaths occurred during the study, during treatment or during the first 3 days after completing therapy and between and 8 weeks after treatment was completed. Eleven deaths were directly related to the non-infectious underlying disease process (eight patients with neurological diseases, including five cerebrovascular accidents and one Guillain-Barre syndrome, one with a myocardial infarction, one with bronchial carcinoma and one with a pulmonary embolism). In four other cases, death was also related to the underlying disease, but the infectious process was a contributing factor (one patient each with heart failure, pancreatic carcinoma, bronchial carcinoma and alcoholic cirrhosis). Finally, seven deaths occurred as a result of septic shock or Table V. Bacteriological response Pathogen Eradication Persistence Methicillin-susceptible S. aureus 2 4 Methicillin-susceptible S. epidermidis S. pneumoniae 2 Streptococcus agalacliae 4 Streptococcus mills Streptococcus (group C) Streptococcus sanguis Escherichia coli 9 K. pneumoniae 4 Klebsiella oxytoca 4 E. cloacae 4 E. agglomerans E. aerogenes 2 2 Serratia spp. Serratia marcescens 2 Proteus vulgar is I Proteus mirabilis Morganella morganii 2 P. aeruginosa 9 3 Acinetobacter baumannii 2 H. influenzae 6 Moraxella catarrhalis 2 Total 7(9%) (9%)

Cefepime and amikacin in ICU patients 2 Table VI. Bacteriological response septicaemia/bacteraemia Organism Eradication Persistence Indefinite Total Septicaemia/bacteraemia associated with respiratory infection Methicillin-susceptible 5. aureus S. pneumoniae E. coli Other septicaemia/bacteraemia'' A. calcoaceticus Total "Patient clinically cured but repeat cultures not obtained. ''Source of infection not determined. multiple organ failure of infectious origin. No aetiological agent was identified in two cases. The other five cases included two with multiresistant P. aeruginosa, two with 5. aureus and one with Enterobacter aerogenes. Six of the early deaths occurred in patients with an initial SAPS of between 5 and 25. The causes of death included cerebrovascular accident (five patients), bronchial carcinoma (two) and multiple organ failure of infectious origin (four), including one caused by methicillin-susceptible S. aureus infection despite the addition of teicoplanin. The incidence of early death was greater (P < 5) in the patients whose SAPS were above 5 (6/3; 2%) than in those with a SAPS below 5 (5/88; 5-7%). The other deaths occurred at least one week after completion of the treatment course. Nine of the 79 evaluable patients had positive blood cultures (in at least two sets in seven cases). These nine patients also had the sepsis syndrome and four met the criteria for septic shock. Nine pathogens were isolated; in six cases, the blood culture isolate was also isolated from the respiratory tract specimen. In two of these patients, the causative pathogen was either resistant to cefepime (methicillin-resistant S. epidermidis) or was not tested and was assumed to be resistant {Bacleroides vulgatus). The bacteriological responses of the other seven patients are shown in Table VI. Eighteen new infections (2 respiratory) were observed either during (2) or after (six) treatment; three were non-bacterial (two during treatment, one post-treatment) (two Candida spp. and one cytomegalovirus infection). The bacterial pathogens isolated during treatment included seven staphylococci (three S. aureus and four S. epidermidis), including five methicillin-resistant strains, one Acinetobacter sp., one cefepime-resistant B. vulgatus, one Enterobacter cloacae and one Enterococcus faecalis. The causes of the five new bacterial infections which presented after completion of treatment included three strains resistant to cefepime {Enterobacter agglomerans, Bacteroides thetaiotaomicron and E. faecalis), two methicillin-resistant strains of S. epidermidis and one Acinetobacter anitratus isolate which was not tested for susceptibility to cefepime. Twelve of these 8 new infections involved the respiratory tract. 2 2 5 o-o 3 2 7 Adverse events None of the 22 deaths was attributed to cefepime or amikacin. Two patients discontinued treatment prematurely due to a clinical adverse event (rash). No other clinical adverse events or local intolerance were reported.

22 F. Gouin et al. Of the patients with normal laboratory results at entry, five patients had six significantly abnormal laboratory test results possibly related to treatment (elevated blood urea (two), serum glutamic-oxaloacetic transaminase, serum glutamic-pyruvic transaminase, creatinine and platelet count). Of the patients who had abnormal laboratory test results at entry, seven exhibited a significant worsening, possibly related to treatment (elevated blood urea (three), alkaline phosphatase (two), creatinine, prothrombin time, platelet count and eosinophil count). However, none of the 8 patients discontinued treatment due to these abnormal test results. Discussion The spectrum of activity of cefepime includes all Enterobacteriaceae, Pseudomonas spp. and methicillin-susceptible staphylococci. Due to its low affinity for /?-lactamases, it retains bactericidal activity in the presence of large inocula. In vivo, the activity of cefepime is equivalent to that of ceftazidime in various models of infection caused by Gram-negative bacteria. Its broad spectrum might therefore make it a suitable choice for the empirical treatment of nosocomial, lower respiratory tract infections. Owing to the severity of these infections, treatment is often based on a combined regimen of a /Mactam antibiotic and an aminoglycoside. The purpose of this study was to evaluate the efficacy and tolerance of cefepime 2 g bd in combination with amikacin in the treatment of infections occurring in ICU, especially nosocomial pneumonia. The severity of the respiratory tract infections in this study was attested to by SAPS, since 75% of the patients had a SAPS of or above on inclusion (mean SAPS, 2). In addition, 84% of the 3 patients had the sepsis syndrome and 35% were in septic shock. Ninety-six per cent of the patients were receiving assisted ventilation. The bacteria isolated were representative of those commonly identified in ICU patients, with S. aureus (9%), P. aeruginosa (4%) and Enterobacteriaceae of the Klebsiella, Enterobacter, Serratia groups (7%) predominating. Of the total of 55 organisms tested for susceptibility to cefepime, 4 (9%) were susceptible and 6 (4%) were moderately susceptible, while (7%) were susceptible and 7 (%) were moderately susceptible to amikacin. Seventy-nine (7%) of the 3 patients with respiratory tract infections were evaluable for efficacy. Of the 34 patients who could not be evaluated, 25 were nonevaluable for reasons unrelated to the choice of therapy. The other nine cases included eight in which a cefepime-resistant pathogen was isolated and one premature discontinuation of amikacin. A satisfactory clinical response was observed in 68 (86%) of the 79 evaluable infections. Of the 7 pathogens isolated from these 79 patients, 7 (9%) were eradicated. The ten bacteriological failures were associated with clinical failure in eight cases and included four strains of S. aureus (including two which were methicillinresistant), two strains of E. aerogenes, three of P. aeruginosa and one of K. pneumoniae. Twenty-two deaths occurred in the course of the study, during treatment or during the first three days thereafter. None of these deaths was attributed to either cefepime or amikacin. Of the early deaths, six patients had an initial SAPS of between 5 and 25. The incidence of early death was higher {P < 5) among those patients whose SAPS were above 5 (2%, 6/3) than in those with a SAPS below 5 (5-7%, 5/88). The other deaths occurred between and 8 weeks after completion of treatment.

Cefepime and amikacin in ICU patients 23 The clinical tolerance of the cefepime-amikacin combination was good. Two adverse events, both of which were rashes, were observed and led to the early discontinuation of treatment. No local intolerance was reported. None of the patients discontinued treatment due to abnormal laboratory test results. In conclusion, cefepime 2 g bd in combination with amikacin was found to be effective and well tolerated in the treatment of nosocomial lower respiratory tract infections in ICU patients, most of whom were being mechanically ventilated. Further comparative controlled studies therefore appear justified in order to determine its role in the empirical management of nosocomial pneumonia. Acknowledgement This investigation was supported by a grant from Bristol-Myers Squibb Company, NJ, USA. References American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference. (992). Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Critical Care Medicine 2, 864-74. Barbhaiya, R. H., Forgue, S. T., Gleason, C. R., Knupp, C. A., Pittman, K. A., Weidler, D. J. el al. (99). Safety, tolerance, and pharmacokinetic evaluation of cefepime after administration of single intravenous doses. Antimicrobial Agents and Chemotherapy 34, 8-22. Barbhaiya, R. H.. Forgue, S. T., Shyu, W. C, Papp, E. A. & Pittman, K. A. (987). High-pressure liquid chromatographic analysis of BMY-2842 in plasma and urine. Antimicrobial Agents and Chemotherapy 3, 55-9. Barbhaiya, R. H., Knupp, C. A. Pfeffer, M. & Pittman, K. A. (992). Lack of pharmacokinetic interaction between cefepime and amikacin in humans. Antimicrobial Agents and Chemotherapy 36, 382-6. Fagon, J. Y., Chastre, J., Domart, Y., Trouillet, J. L., Pierre, J., Darne, C. el al. (989). Nosocomial pneumonia in patients receiving continuous mechanical ventilation. Prospective analysis of 52 episodes with use of a protected specimen brush and quantitative culture techniques. American Review of Respiratory Disease 39, 877-84. Fung-Tome, J., Dougherty, T. J., DeOrio, F. j., Simich-Jacobson, V. & Kessler, R. E. (989). Activity of cefepime against ceftazidime- and cefotaxime-resistant Gram-negative bacteria and its relationship to /Mactamase levels. Antimicrobial Agents and Chemotherapy 33, 498-52. Fung-Tome, J., Huczko, E., Pearce, M. & Kessler, R. E. (988). Frequency of in vitro resistance of Pseudomonas aeruginosa to cefepime, ceftazidime and cefotaxime. Antimicrobial Agents and Chemotherapy 32, 443-5. Kessler, R. E., Bies, M., Buck, R. E., Chisholm, D. R., Pursiano, T. A., Tsai, Y. H. et al. (985). Comparison of a new cephalosporin, BMY 2842, with other broad-spectrum /Mactam antibiotics. Antimicrobial Agents and Chemotherapy 27, 27-6. Marchou, B., Michea-Hamzehpour, M., Lucain, C. & Pechere, J.-C. (987). Development of /Mactam-resistant Enterobacter cloacae in mice. Journal of Infectious Diseases 56, 369-73. Masuyoshi, S., Hiraoka, M., Inoue, M., Tomatsu, K., Hirano, M. & Mitsuhashi, S. (989). Comparison of the in vitro and in vivo antibacterial activities of cefepime (BMY-2842) with ceftazidime, cefuzonam, cefotaxime and cefmenoxime. Drugs under Experimental and Clinical Research 5, -. Papazian, L., Martin, C, Meric, B., Dumon, J. F. & Gouin, F. (993). A reappraisal of blind bronchial sampling in the microbiologic diagnosis of nosocomial bronchopneumonia. A comparative study in ventilated patients. Chest 3, 236-42. Phelps, D. J., Carlton, D. D., Farrel, C. A. & Kessler, R. E. (986). Affinity of cephalosporins

24 F. Gouin et al. for /i-lactamases as a factor in antibacterial efficacy. Antimicrobial Agents and Chemotherapy 29, 845-8. Rouby, J. J.. Martin de Lassale. E.. Poete. P.. Nicolas, M. H., Bodin. L. Jarlier. V. el al. (992). Nosocomial bronchopneumonia in the critically ill. Histologic and bactenologic aspects. American Review of Respiratory Disease 46, 59-66. Tomatsu, K.., Ando, S., Masuyoshi, S., Hirano. M., Miyaki. T. & Kawaguchi, H. (986). Antibacterial activity of BMY-2842, a novel broad-spectrum cephalosponn. Journal of Antibiotics 39, 584-9.