The role of Gram-positive therapy in the neutropenic patient

Transcription

1 Journal of Antimicrobial Chemotherapy (1991) 27, Suppl. B, The role of Gram-positive therapy in the neutropenic patient Francesco Menichetti" and Albano Del Favero* "Institute of Infectious Diseases; b Department of Internal Medicine, Perugia University, Perugia, Italy The increasing prevalence of Gram-positive infections in neutropenic cancer patients seems to be related to the use of central venous catheters, chemotherapy-induced oral and gastrointestinal mucositis, and the prophylactic use of fluoroquinolones. The need for anti-gram-positive therapy in the neutropenic patient is supported by the increasing prevalence and the changing resistance of Gram-positive pathogens, as well as by the poor response of Gram-positive bacteraemia to aminoglycoside plus /Mactam regimens. Combined therapy with either vancomycin or and other empirical antibiotics, has proved efficacious in adults and children with neutropenia, fever and Gram-positive infection. Vancomycin exerts greater antibacterial activity against strains of coagulase-negative staphylococci than and there is more data on its routine clinical use. In its favour, is less toxic and easier to administer. The time when a glycopeptide antibiotic should be introduced is still a matter of debate; support for both initial therapy and subsequent rescue therapy is found in the current literature. Large clinical trials are warranted to clarify further the role of anti-gram-positive therapy in the neutropenic patient. Introduction The increasing prevalence of Gram-positive infections noted in recent years in neutropenic cancer patients in many centres and clinical trials suggest that an anti-grampositive antibiotic should be included in empirical therapy regimens for febrile, neutropenic patients (Del Favero et al., 1988). Although many clinical trials have been conducted to evaluate the role of anti-gram-positive therapy in the neutropenic patient, opinions remain divided on the need for therapy, the choice of the antibiotic to be used and the timing for introducing the antibiotic in the empirical regimen. The need for anti-gram-positive therapy in neutropenic patient The main reason for considering anti-gram-positive therapy in the management of the febrile neutropenic patient is the changing epidemiological pattern of infections in these patients. The five therapeutic trials carried out between 1973 and 1988 by the EORTC International Antimicrobial Therapy Cooperative Group showed that the rate of single-agent Gram-negative bacteraemias decreased from 71% to 36% and that of single-agent Gram-positive bacteraemias increased from 29% to 64%. Possible explanations for the present prevalence of Gram-positive infections in neutropenic Correspondence to: Dr Francesco Menichetti, Istituto di Malattie Infettive, Universita' di Perugia, Ospedale Policlinico Regionale, Perugia, Italy /91/27BO $02.00/ The British Society for Antimicrobial Chemotherapy

2 52 F. Menichetti and A. Del Favero patients are the increasing use of central venous catheters (CVC) and the severe, chemotherapy-induced, oral and gastrointestinal mucositis, both of which are portals of entry for the subsequent haematogenous spread of colonizing micro-organisms (Viscoli, Van der Auwera & Meunier, 1988). Prophylactic fluoroquinolones are probably another important risk factor, since, although they are able to prevent Gramnegative infections, they are less efficacious against those caused by Gram-positive micro-organisms (Young, 1987). In the GIMEMA cooperative study, among 619 neutropenic patients receiving bacterial prophylaxis with norfloxacin or ciprofloxacin, Gram-positive bacteraemias were more frequent (61/95, 64%) and the pathogens most isolated were coagulase-negative staphylococci and streptococci, which respectively accounted for 28% and 26% of all blood isolates (GIMEMA, 1989). These Grampositive cocci may be poorly susceptible to the antibiotics currently used for the empirical therapy of the neutropenic patient, such as a /Mactam plus an aminoglycoside. All 44 Gram-positive cocci which were isolated from blood of febrile neutropenic patients empirically treated with amikacin plus ceftazidime and proved to be susceptible to, whereas 59% were resistant to ceftazidime and 48% to amikacin. In particular, 67% of coagulase-negative staphylococci proved to be resistant to ceftazidime and 42% to amikacin; 85% of streptococci were shown to be resistant to amikacin and 46% to ceftazidime. Because of this susceptibility pattern, was the only administered antibiotic that was active against the infecting strain in vitro in 30% of Gram-positive bacteraemias (Menichetti etal., 1990). Further support for the specific need for anti-gram-positive therapy in neutropenic patients, is the observation of a poor response of Gram-positive bacteraemias to various combinations of aminoglycoside and /Mactam antibiotics. In EORTC therapeutic trial FV, less than 50% of the neutropenic patients with Gram-positive bacteraemias improved with the combination of amikacin (for three or nine, days) and azlocillin or ceftazidime (EORTC, 1987). All these data indicate the need for anti-gram-positive therapy and, particularly, for a glycopeptide antibiotic, in the management of febrile neutropenic cancer patients. The use of glycopeptide antibiotics in the neutropenic patient Vancomycin has been extensively tested in neutropenic patients and shown to be useful both in initial therapy (Karp etal., 1986; Kramer, Ramphal & Rand, 1986; Shenep et al., 1988) and in rescue therapy (Pizzo et al., 1986; Granowetten, Wells & Lange, 1988). Two randomized, double-blind, placebo-controlled studies, one in children (Shenep et al., 1988) and one in adults (Karp et al., 1986) found that febrile neutropenic cancer patients, who initially received vancomycin in addition to an aminoglycoside and an antipseudomonal penicillin had more rapid resolution of the initial febrile episode, fewer days of bacteraemia, and a lower incidence of treatment failure than those who did not receive vancomycin. Kramer et al. (1986) reported initial treatment with ceftazidime plus vancomycin to be superior to that with ceftazidime alone or with cephalothin, gentamicin, and carbenicillin. In contrast, the EORTC therapeutic trial V, which compared amikacin plus ceftazidime with and without initial vancomycin, failed to demonstrate any advantage of the regimen that included the glycopeptide antibiotic. In fact, although single Gram-positive bacteraemias responded apparently better to the three-drug combination, the proportion of febrile patients on each trial day and the

3 Glycopeptide antibiotics in the neutropenic patient 53 Table I. Teicoplanin in initial empirical combined therapy Study design: patient population Open: 46 adults, BMT (Menichetti et al, 1988) Open: 46 AL children (Schaison et al., 1988) RCT: 66 AL adults (Del Favero et al, 1987) RCT: 108 adults with HM (Novakova et al, 1988a) Therapeutic regimen ( dosage) amikacin/ ceftazidime/ (5 mg/kg/day) ceftriaxone/ (10 mg/kg/day) amikacin/ ceftazidime amikacin/ ceftazidime/ (5 mg/kg/day) ceftazidime ceftazidime/ (5 mg/kg/day) Response to therapy in Gram-positive cocci bacteraemias (patients) 18/20 (20%) 41/46 (89%) 1/4 (25%) 4/5 (80%) 7/18 (39%)* 16/21 (76%)* Sideeffects 1 rash 1 hypoacuesthesia 1 increase of creatinine none none none Notes six strains susceptible to only patients with neutrophil count < 100/mm 3 30% 83% AL, Acute leukaemia; BMT, bone marrow transplantation; HM, haematological malignancies; RCT, randomized comparative trial. "Overall response. b P< duration of fever were similar in the two treatment groups and patients treated with vancomycin had a higher nephrotoxicity rate (EORTC, 1988). In other studies, febrile neutropenic patients received either ceftazidime alone or other antibiotics as initial empirical therapy, and vancomycin was used as needed later in the course of treatment (Pizzo et al, 1986; Granowetten et al, 1988). There was no increase in mortality or morbidity rates due to delaying starting vancomycin therapy until deemed necessary. Thus, when included in initial therapy and when used as rescue therapy in non-responders, vancomycin seems to be useful for the management of the febrile, neutropenic patient. Teicoplanin, a glycopeptide antibiotic with antibacterial activity similar to vancomycin and a pharmacokinetic profile which allows once-daily administration, has also been evaluated in the neutropenic patient. When used in the initial antibiotic therapy (Table I), in open trials, it was shown to be safe and effective for the management of Gram-positive bacteraemias both in adult patients undergoing bone-marrow transplantation (Menichetti et al, 1988) and in children with acute leukaemia (Schaison etal, 1988). Two randomized controlled studies which added to ceftazidime (Novakova etal, 1988a) or combination of amikacin and ceftazidime (Del Favero etal, 1987) demonstrated that these regimens were more effective for the treatment of Gram-positive bacteraemias in adult patients with haematological malignancies than comparative regimens which did not include the glycopeptide. An amikacin plus ceftazidime and regimen gave notably better results in patients with profound (neutrophil count < 100/mm 3 ) and persistent

4 54 F. Menichetti and A. Del Favero Table II. Teicoplanin in rescue therapy Study design: patient population Open: 52 adults with HM (Novakova et al., 1988") Open: 20 children with cancer (15 neutropenic) (Lemerle et al., 1988) Therapeutic regimen and dosage ceftazidime in non-responders: 20 patients 200 mg/day 32 patients 400 mg/day aminoglycoside plus /J-lactam in nonresponders: 6-10mg/kg/day HM, Haematological malignancies. "Overall response. Study design: patient population Open: 19 cancer adults with CVC (16neutropenics) coagulase-negative staphylococci infections RCT: (Webster et al, 1987) 60 CVC infections in cancer patients (51 neutropenics) (Smith et al., 1989) Side- effects Percentage response to therapy in Gram-positive cocci bacteraemia 73%" 2 increased SGOT/SGPT 67% 4 increased SGOT/SGPT 2 hypo-acousia 73% (8/11) 8 in 6 patients Table III. Teicoplanin in Central Venous Catheter (CVC) Infection Therapeutic regimen and dosage gentamicin/ cefuroxime/ 3 mg/kg/day gentamicin/ piperacillin/ vancomycin gentamicin/ piperacillin/ 11 patients 200 mg/day 49 patients 400 mg/day RCT, Randomized comparative trial. Response to therapy sepsis exit-site tunnel Notes 7/9 9/10 CVC removal: 3/19 6/6 10/10 0/4 4 mucocutanous reactions 5 nephrotoxicity 8/8 7/11 3/7 lrash 1 nephrotoxicity neutropenia (Del Favero et al., 1987). There were no differences in adverse effects of the antibiotic regimens. As rescue therapy in neutropenic patients not responsive to the initial empirical antibiotic regimen (Table II), was effective in adult patients with haematological malignancies unresponsive to ceftazidime (Novakova et al., 1988A) and children with cancer unresponsive to aminoglycoside plus /Mactam (Lemerle etal., 1988).

5 Glycopeptide antibiotics in the neutropenic patient 55 Table IV. Teicoplanin in elective therapy of Gram-positive infections Study design: patient population RCT: 40 cancer adults (Gerard et al., 1987) RCT: 34 cancer adults (Aoun et al., 1989) Therapeutic regimen and dosages vancomycin 1 g bid 0 400mgqd(3days) 200 ing qd vancomycin (HP) lgbid * 400mgtid (1 day) 400 mg qd Response to therapy in Gram-positive cocci bacteraemias (patients) 9/11 8/10 11/15 7/10 HP, highly purified; RCT, randomized comparative trial. 400 mg od for three days followed by 200 mg od. '400 mg tid for first day followed by 400 mg od. Peak serum levels (mg/1) Side effects 4 rash 4 superinfection 1 increased creatinine level 1 superinfection 1 superinfection 1 increased creatinine level 1 superinfection Teicoplanin was also of value in treating central venous catheter (CVC)-related infections (Table III). In an open trial carried out on cancer patients with infections caused by coagulase-negative staphylococci, an empirical antibiotic regimen including gentamicin, cefuroxime and proved to be highly effective, as only 3/19 infected patients required catheter removal (Webster et al., 1987). A prospective, controlled trial on patients with haematological malignancies and CVC-related infections randomized to receive gentamicin plus piperacillin combined with vancomycin or found both regimens to be equally effective, but -treated patients had fewer side-effects (Smith et al., 1989). Teicoplanin has been compared with vancomycin in the elective therapy of documented Gram-positive infections in adult cancer patients (Table IV). Two consecutive but small randomized, comparative trials, one with 200 mg/day (Gerard et al., 1987) the other with 400 mg/day (Aoun etai, 1989), found to be equally effective as vancomycin, and the higher dose provoked no increase in related toxicity. The choice between the glycopeptide antibiotics Comparative analysis of the characteristics of the two glycopeptides shows that both compounds have similar antibacterial activity, except for some strains of coagulasenegative staphylococci (Staphylococcus haemolyticus and S. epidermidis), which are more susceptible to vancomycin (Goldstein et al., 1990). It is noteworthy that previous treatment with a glycopeptide may be responsible for the onset of resistance to in coagulase-negative staphylococci, as has been shown in strains isolated from neutropenic patients (Maugein et al, 1990).

6 56 F. Menicherti and A. Del Favero Both glycopeptide antibiotics are synergistic l in vitro' with, for example, aminoglycosides, that are usually a component of empirical antibiotic regimens. Although greater clinical experience has been gained with vancomycin, the efficacy of the two glycopeptides in the combined therapy of Gram-rpositive infections in neutropenic patients has been demonstrated to be similar. However, clinical failure of therapy in neutropenic patients with staphylococcal bacteraemia, who recovered after administration of vancomycin, has been reported (Brunet etal., 1990). Vancomycin is potentially nephrotoxic; factors found to be associated with increased risk were concurrent therapy with an aminoglycoside, length of treatment with vancomycin above 21 days, and vancomycin trough serum concentration above 10mg/l (Rybak et ah, 1990). Therefore, in neutropenic patients treated with vancomycin and an aminoglycoside, it seems advisable to monitor serum levels of vancomycin closely. Teicoplanin seems to be less toxic. Comparison of the nephrotoxic potential of the two glycopeptides in neutropenic patients treated with piperacillin, tobramycin and or vancomycin, revealed that nephrotoxicity occurred only in patients who recieved vancomycin (Kureishi etal., 1989). Furthermore, despite the fact that is administered by rapid (3-4 min) iv injection, hypotension rarely occurs and the characteristic 'red-man syndrome' described with vancomycin is seldom encountered. A prospective comparison of vancomycin- and -induced histamine release and 'red man syndrome' reported that vancomycin caused the syndrome in 92% of the patients and was also associated with a significant increase in plasma histamine; in contrast, neither caused 'red-man syndrome' nor elicited significant histamine release (Sahai etal., 1990). It has been suggested that an infusion of purified vancomycin over 1 h at a concentration of no more than 5 g/1 circumvents the 'red-man syndrome', at least in neutropenic children (Shenep etal., 1988). Teicoplanin has the advantage of being much easier to administer; only one daily dose is required against the two to four, each given over 60 min, needed for vancomycin. In consequence, it saves valuable nursing time. The choice of the most appropriate dose and administration schedule for is of crucial importance. There seems to be a relationship between the dose, the peak serum levels and the clinical effectiveness of. Most reported cases of treatment failures (Calain et al, 1987) are clearly related to insufficient dosage (i.e. 200 mg/daily). Increasing the dose of administered to patients with infective endocarditis to 7-14mg/kg/day should guarantee higher serum levels and better therapeutic response (Martino etal., 1989). In our opinion, the higher dosages (i.e. lomg/kg/day) should be reserved for monotherapy. An appropriate administration schedule for in combined therapy would be 6mg/kg, administered 12-hourly for the first three doses, then once daily. Further large-scale clinical trials that compare the value of the two glycopeptides are essential to establish clearly which is the better choice. The timing for the introduction of glycopeptide antibiotics There is still controversy over the exact time glycopeptide antibiotics should be introduced into therapeutic regimens for the management of neutropenic patients. Table V lists the factors in support of both initial empirical and rescue therapy. The use of a glycopeptide antibiotic as first-line therapy is not only supported by epidemiological considerations but also by the better results obtained in neutropenic patients

7 Glycopeptide antibiotics in the neutropenic patient 57 Table V. Factors favouring the use of glycopeptide antibiotic in initial and rescue therapy Initial therapy Rescue therapy Prevalence of Gram-positive cocci infections High incidence of Gram-positive cocci resistant to aminoglycoside and/or 0-lactams Poor response of Gram-positive cocci bacteraemias to aminoglycoside plus /Mactams Better response when vancomycin is initially included (toxicity implications) Better response when is initially included (low toxicity) Prevention of superinfections or secondary infections by Gram-positive cocci Possible reduction of amphotericin use Low incidence (10%) of Gram-positive cocci infections in neutropenic febrile episodes Good activity of ureidopenicillin against streptococci Low mortality rate with Gram-positive infections Rescue therapy always successful Limitation in overuse of glycopeptides (less cost, and lower toxicity and risk of resistance) Possible reduction of amphotericin use when vancomycin or have been added initially to the empirical antibiotic regimen. Furthermore, there are data suggesting that the initial use of a glycopeptide antibiotic reduces superinfections or subsequent infections caused by Gram-positive organisms (Karp el ah, 1986; Shenep etal., 1988) and reduces the need for empirical amphotericin B (Karp etal, 1986). The factors that favour reserving glycopeptide antibiotics for second-line therapy are the overall low rate of Gram-positive infections in neutropenic febrile episodes, which now account for no more than 10-15% of all febrile episodes, the low early mortality rate of these infections and the good results obtained with rescue therapy. Therefore, it seems feasible to delay therapy until Grampositive infections have been microbiologically documented. Such an approach would limit the cost, toxicity and the potential risk of the emergence of resistant strains due to overuse of glycopeptide antibiotics. It is noteworthy that a reduction of amphotericin B use has also been noted in patients receiving vancomycin as rescue therapy (EORTC, 1988). Both approaches, therefore, find their rationale in current literature. In our opinion, however, it seems wise to include a glycopeptide antibiotic in the initial empirical regimen of neutropenic patients receiving bacterial prophylaxis with fluoroquinolones, who have indwelling CVC and suffer from severe oral and gastrointestinal mucositis; all common features in patients affected by acute leukaemia or undergoing bone marrow transplantation. Further studies are needed to identify clearly the subgroup of neutropenic patients at highest risk of Gram-positive infections, who would benefit from the initial use of a glycopeptide antibiotic. Conclusion The increasing prevalence and changing resistance pattern of Gram-positive pathogens, as well as their poor response to an aminoglycoside plus /Mactam regimen point to the need for specific anti-gram-positive therapy in the treatment of the neutropenic

8 58 F. Menichetti and A. Del Favero patient. Both vancomycin and have proved effective; vancomycin possesses a greater antibacterial activity against coagulase-negative staphylococci, but is less toxic and easier to administer. The timing for the introduction of a glycopeptide antibiotic is still controversial. Further clinical trials are needed to clarify the role of anti-gram-positive therapy in the neutropenic patient. References Aoun, M., Van Der Auwera, P., Delhaye, N. & Meunier, F. (1989). Prospective, randomized study of vancomycin and for severe Gram-positive infections in immunocompromised patients. In Program and Abstracts of the 29th Interscience Conference on Antimicrobial Agents and Chemotherapy, Houston, Abstract 596. American Society for Microbiology, Washington, DC. Brunet, F., Vedel, G., Dreyfus, F., Vaxelaire, J. F., Giraud, T., Schremmer B. etal. (1990). Failure of therapy in two neutropenic patients with staphylococcal septicemia who recovered after administration of vancomycin. European Journal of Clinical Microbiology and Infectious Diseases 9, Calain, P., Krause, K. H., Vaudaux, P., Auchentaler, R., Lew, D., Waldvogel, F. etal. (1987). Early termination of a prospective, randomized trial comparing and flucloxacillin for treating severe staphylococcal infections. Journal of Infectious Diseases 155, Del Favero, A., Menichetti, F., Guerciolini, R., Bucaneve, G., Baldelli, F., Aversa, F. etal. (1987). Prospective randomized clinical trial of for empiric combined antibiotic therapy in febrile, granulocytopenic acute leukemia patients. Antimicrobial Agents and Chemotherapy 31, Del Favero, A., Menichetti, F., Bucaneve, G., Minotti, V. & Pauluzzi, S. (1988). Septicaemia due to Gram-positive cocci in cancer patients. Journal of Antimicrobial Chemotherapy 21, Suppl. C, EORTC International Antimicrobial Therapy Project Group. (1987). Ceftazidime combined with a short or long course of amikacin for empirical therapy of Gram-negative bacteremia in cancer patients with granulocytopenia. New England Journal of Medicine 317, EORTC International Antimicrobial Therapy Project Group (1988). Ceftazidime plus amikacin with or without vancomycin as empirical therapy of fever in cancer patients with granulocytopenia. In Program and Abstracts of the 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, Abstract 18. American Society for Microbiology, Washington, DC. Goldstein, F. W., Coutrot, A., Sieffer, A. & Acar, J. (1990). Percentages and distribution of - and vancomycin-resistant strains among coagulase-negative staphylococci. Antimicrobial Agents and Chemotherapy 34, Gerard, M., Van Der Auwera, P., Meunier, F., Ninove, D., Daneau, D. & KJastersky, J. (1987). A controlled clinical trial on efficacy and safety of versus vancomycin in the treatment of staphylococcal infections. In Program and Abstracts of the 27th Interscience Conference on Antimicrobial Agents and Chemotherapy, New York, Abstract American Society for Microbiology, Washington, DC. Granowetten, L., Wells, H. & Lange, B. J. (1988). Ceftazidime with or without vancomycin. cephalothin, carbenicillin and gentamicin as initial therapy of the febrile neutropenic pediatric cancer patient. Pediatric Infectious Diseases Journal 7, Gruppo Italiano Malattie Ematologiche Maligne dell'adulto (GIMEMA) (1989). Comparison of norfloxacin with ciprofloxacin for infection prophylaxis in neutropenic patients with acute leukemia. A randomized, prospective, multicenter study. In Program and Abstracts of the 29th Interscience Conference on Antimicrobial Agents and Chemotherapy, Houston, Abstract 600. American Society for Microbiology, Washington, DC. Karp, J. E., Dick, J. D., Angelopulos, C, Charache, P., Green, L., Burke, P. J. etal. (1986). Empiric use of vancomycin during prolonged treatment-induced granulocytopenia. American Journal of Medicine 81,

9 Glycopeptide antibiotics in the nentropenic patient 59 Kramer, B. S., Ramphal, R. & Rand, K. H. (1986). Randomized comparison between two ceftazidime-containing regimens and cephalothin-gentamicin-carbeniciuin in febrile granulocytopenic cancer patients. Antimicrobial Agents and Chemotherapy 30, Kureishi, A., Jewesson, P. J., Cole, C. D., Reece, D. E., Phillips, G. L., Smith, J. A. etal. (1989). The nephrotoxic potential of. vancomycin in the empiric treatment of febrile neutropenic patients. In Program and Abstracts of the 29th Interscience Conference on Antimicrobial Agents and Chemotherapy. Houston, Abstract 597. American Society for Microbiology, Washington, DC. Lemerle, S., de La Rocque, F., Lamy, R., Fremaux, A., Bernaudin, F., Lobut, J. B. etal. (1988). Teicoplanin in combination therapy for febrile episodes in neutropenic and non-neutropenic paediatric patients. Journal of Antimicrobial Chemotherapy 21, Suppl. A, Martino, P., Venditti, M., Micozzi, A., Brandimarte, C, Gentile, G., Santini, C. etal. (1989). Teicoplanin in the treatment of Gram-positive bacterial endocarditis. Antimicrobial Agents and Chemotherapy 33, Maugein, J., Pellegrin, J. L., Brossard, G., Fourche, J., Leng, B. & Reiffers, J. (1990). In vitro activities of vancomycin and against coagulase-negative staphylococci isolated from neutropenic patients. Antimicrobial Agents and Chemotherapy 34, Menichetti, F., Del Favero, A., Bucaneve, G., Aversa, F., Baldelli, F., Felicini, R. etal. (1988). Teicoplanin in empirical combined antibiotic therapy of bacteremias in bone marrow transplant patients. Journal of Antimicrobial Chemotherapy 21, Suppl. A, Menichetti, F., Del Favero, A., Bucaneve, G., Fiorio, M. & Aversa, F. (1990). Using for empiric therapy of febrile neutropenic patients with haematologjcal malignancies. British Journal of Haematology, 76, Suppl. 2, Novakova, I., Donnelly, J. P., Muytjens, H. L. & De Pauw, B. E. (1988a). Ceftazidime with and without as initial empirical therapy for febrile granulocytopenic patients. In Program and Abstracts of the 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, Abstract 16. American Society for Microbiology, Washington, DC. Novakova, I. R. O., De Pauw, B. E., Verhagen, S., Muytjens, H. & Donnelly, J. P. (1988*). Teicoplanin rescue for Gram-positive ceftazidime failures during neutropenia. Abstracts of the 5th International Symposium on Infections in the Immunocompromised Host, Noordwijkerhout, The Netherlands, Abstract 35. Pizzo, P. A., Hatorn, J. W., Hiemenz, J., Brown, M., Cotton, D., Gress, J. etal. (1986). A randomized trial comparing ceftazidime alone with combination antibiotic therapy in cancer patients with fever and neutropenia. New England Journal of Medicine 315, Rybak, M. J., Albrecht, L. M., Boike, S. C. & Chandrasekar, P. H. (1990). Nephrotoxicity of vancomycin, alone and with an aminoglycoside. Journal of Antimicrobial Chemotherapy 25, Sahai, J., Healy, D. P., Shelton, M. J., Miller, J. S., Ruberg, S. J. & Polk, R. (1990). Comparison of vancomycin- and -induced histamine release and 'Red Man Syndrome'. Antimicrobial Agents and Chemotherapy 34, Schaison, G., Leverger, G., Arlet, G., Pecking, M. & Garaud, J. J. (1988). A once a day antimicrobial therapy: and ceftriaxone in the treatment of febrile neutropenic children with acute leukemia. In Program and Abstracts of the 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, Abstract 19. American Society for Microbiology, Washington, DC. Shenep, J., Hughes, W. T., Roberson, P. K., Blankenship, K.., Baker, D. & Meyer, W. H. (1988). Vancomycin, ticarcillin and amikacin compared with ticarcillin-clavulanate and amikacin in the empirical treatment of febrile neutropenic children with cancer. New England Journal of Medicine 319, Smith, S. R., Cheesbrough, J., Spearing, R. & Davies, J. (1989). Randomized prospective study comparing vancomycin with in the treatment of infections associated with Hickman catheters. Antimicrobial Agents and Chemotherapy 33, Viscoli, C, Van der Auwera, P. & Meunier, F. (1988). Gram-positive infections in granulocytopenic patients: an important issue? Journal of Antimicrobial Chemotherapy 21, Suppl. C,

10 60 F. Menichetti and A. Del Favero Webster, A., Russell, S. J., Souhami, R. L., Richards, J. D. M., Goldstone, A. H. & Gruneberg, R. N. (1987). Use of for Hickman catheter associated staphylococcal infection in immunosuppressed patients. Journal of Hospital Infection 10, Young, L. S. (1987). The new fluorinated quinolones for infection prevention in acute leukemia. Annals of Internal Medicine 106,