British Journal of Huematology. 1990. 76, Suppl. 2, 10-13

A randomized study of teicoplanin plus ciprofloxacin versus gentamicin plus piperacillin for the empirical treatment of fever in neutropenic patients M.K E L S E Y , P E T E R w. C O L L I N S , C O R R I N E DELORD,B A R B A R AW E I N H A R DA*N D A D R I A NC . N E W L A N Departments D of Haernatology and *Microbiology, The London Hospital, Whitechapel, Loridorz STEPHEN

Summary. We compared the combination of teicoplanin plus ciprofloxacin with gentamicin plus piperacillin for the empirical treatment of fever in 8 0 neutropenic patients. A favourable clinical response rate was achieved in 28/38 (74%) patients receiving teicoplanin plus ciprofloxacin and in 17,' 3 5 (49%) of those receiving gentamicin plus piperacillin ( P = 0.05). Microbiologically documented infections accounted for 55% of febrile events. When these episodes were analysed separately, response to teicoplanin plus ciprofloxacin remained unchanged at 74% whereas only 35% patients responded to gentamicin and piperacilin ( P = 0,034). Gram-positive organisms accounted for 78% bacterial isolates with Staphglorocrus epidermidis the most common pathogen. Ten out of 1 2 (83%) Staph. epidermidis

infections resolved when treated with teicoplanin and ciprofloxacin as compared with a response rate of only two out of eight (75%) with gentamicin and piperacillin ( P = O . O 3 2 ) . The combination of teicoplanin and ciprofloxacin was associated with no severe drug-related adverse events: by contrast, two patients receiving gentamicin plus piperacilin were withdrawn owing to adverse drug reactions, one with acute renal failure and one following a severe allergic reaction to piperacillin. We conclude that teicoplanin with ciprofloxacin is more effective than gentamicin plus piperacillin for the empirical treatment of febrile neutropenic patients. The high incidence of Gram-positive infection in our unit probably justifies the use of a specific anti-Gram-positive agent in the first-line antibiotic regimen.

The prevalence of infection with Gram-positive organisms has increased in neutropenic patients over the last 10 years, predominantly owing to selective gut decontamination measures and the use of long-term indwelling central venous catheters in patients receiving chemotherapy for haematological malignancy. Morbidity and mortality from these organisms, particularly coagulase-negative staphylococci, has concomitantly increased (Klastersky, 1986). Teicoplanin is a new glycopeptide antibiotic with specific anti-Gram-positive activity and potential pharmacokinetic and toxicological advantages over vancomycin (Smith et al, 1989; Buniva et al, 1988). In vitro teicoplanin acts additively with ciprofloxacin against staphylococci (C. S. Lewin, unpublished observations). Ciprofloxacin is a broad-spectrum antibiotic with activity against Gram-negative bacteria which is often superior to that of aminoglycosides or cephalosporins (Schiff P t d.1984). We therefore conducted a prospective randomized study comparing the combination of teicoplanin plus ciprofloxacin with our standard regimen of gentamicin plus piperacillin for the empirical treatment of fever in neutropenic patients undergoing bone marrow transplantation or chemotherapy for haematological malignancy. Clinical and microbiological

data on the efficacy of teicoplanin in combination with ciprofloxacin are presented. PATIENTS AND METHODS Eighty neutropenic patients receiving chemotherapy or bone marrow transplant for haematological malignancy were entered into the study between September 1988 and December 1989. All patients were febrile as defined by a single temperature of 38.5"C or two consecutive readings of 38.0"C 2 h apart not associated with administration of blood products. Neutropenia was defined as a total neutrophil count of less than 1.O x 10y/l.All patimts were aged greater than 18 years and gave informed consent. Pregnant or nursing women, patients with a history of allergy to trial antibiotics, and patients who had received antibiotics within the preceding 5 d other than gut decontamination, were excluded from the study. Eligible patients were randomized to receive either teicoplanin plus ciprofloxacin or gentamicin plus piperacillin. Teicoplanin was given as a slow intravenous bolus at a dose of 400 mg 12-hourly for 24 h followed by 400 mg daily. Patients weighing over 100 kg received 6 0 0 mg daily while

10

Randomized Treatment of Fever those with renal impairment (creatinine greater than 170 pmol/l) received teicoplanin on alternate days only after the first 4 d. Ciprofloxacin was given intravenously 2 0 0 mg 12hourly as a 100 ml infusion over 3 0 min. A 1 2 0 mg loading dose of gentamicin was followed by 80 mg 8-hourly to achieve pre-dose troughs of < 2 mg/l and peaks of > 5 mg/l. Piperacillin was administered 4 g 6-hourly as a 50 ml infusion. All patients received gut decontamination with colistin 1 . 5 megaunits b.d. and co-trimoxazole 960 mg twice daily as well as a n oral antifungal agent. At entry to the study full blood count, biochemical profile and chest X-ray were performed; blood analysis was performed at least three times weekly for the duration of the study. Microbiological analysis was performed on peripheral vein and Hickman line-derived blood. urine, throat swab and any other appropriate specimens. Clinical evaluation was made after 2 4 and 72 h of study therapy and at discontinuation of study therapy. Antibiotics were continued until patients had been apyrexial for at least 3 full days. Empirical therapy was deemed to have failed if antibiotics had to be changed or modified at evaluation. including the addition of intravenous amphotericin €3. ModiBcation was performed if: ( a ) a pathogen was isolated which was resistant to both study antibiotics; (b)deterioration in the patient's clinical state before 72 11 was thought to be due tci bactcbrial infection: ( c ) no significant clinical improvement was seen at 72 h; ( d )study drugs were withdrawn owing to toxicity or severe adverse reaction. Clinical resolution of febrile episodes was defined as disapprarance of all signs and symptoms related to the infection while on unmodified study therapy. Improvement was defined as marked reduction in symptoms and signs of infection without complete resolution, or temporary resolution ( > 48 h ) with subsequent failure to maintain this. Any patient who was found to have violated study protocol was not evaluated for clinical efficacy: in these cases clinical outcome was deemed indeterminate but patient data were included in evaluation of drug toxicity. Febrile episodes were classified as microbiologically documented, clinically documented or pyrexia of unknown origin (PUOj in accordance with the recommendations of the Immunoconipromised Host Society (P. A. Pizzo, verbal communication). Microbiologically documented infections were evaluated at the end of therapy and recorded as eradicated. persistent or indeterminate if follow-up cultures could riot be performed. Superinfections were defined as the presence of a new organism during or at the end of study therapy judged to be causing an infectious process.

KESULTS Eighty patients were randomized of which 73 (91%) were evaluable for efficacy. Seven patients were unevaluable: six breached study protocol (four not neutropenic. two inappropriate modification of antibiotics before 72 h ) and one withdrew consent at 2 4 h. Thirty-eight patients received teicoplanin with ciprofloxacin and 3 5 patients gentamicin plus piperacillin. Mean age was 3 4 & 1 3 years for patients receiving teicoplanin plus ciprofloxacin and 39 & 14 years for those receiving gentamicin plus piperacillin. There was no

11

significant difference in presenting diagnoses between the two groups. Mean duration of therapy was 7 i3 d for the teicoplanin group and 5 f 2 d for the gentamicin/piperacillin group. There were 40 microbiologically documented infections (55%). 10 clinically documented infections (14%) and 2 3 PUO (31%). The nature of the documented infections is shown in Table I: 37 of the 40 (93%) microbiologically documented infections were bacteraemia, comprising 74%)of documented infections overall. Twenty-three (61%) febrile episodes occurring in patients randomized to teicoplanin plus ciprofloxacin resolved on therapy and a further five (13%) improved. In the patient group receiving gentamicin plus piperacillin. 1 5 (43%) of febrile episodes resolved and two (6%)improved ( P= 0.0 5. Chi-squared: Table II). There were 4 3 microbial isolates from the 4 0 microbiologically documented febrile episodes. 2 3 occurred in the tecjcoplanin/ciprofloxacin group including one Candida tropiralis. one Asporgillus pnvurnonia and one cytomegalovirus pneumonitis. 20 further pathogens were isolated from the patients receiving gentamicin plus piperacillin. Staph. rpidermidis was the most coninion isolate in both groups accounting for 20 (4770)ofprimary infecting pathogens. Overall. Gram-positive bacterial pathogens organisms comprised 3 1 of the 40 (is'$(?) isolated. These were divided between the treatment group seen in 'I'able Ill. Eradication of the primary pathogen occurred in 13 (61%) of the 23 microbiologically documented episodes in the teicoplaniii/ciprotloxacin group. Four ( 1 7'Xl) organisms persisted (Stoph. c@r/c,rttiit/is. Entvrobnctc~ dnncw. C. tropiccrlis and a cytomegalovirus). Nine ( 5 3%)) of Table 1. Infection sites ofdocumented infections

Kacteraemia Chest Hickman line entry site Throat Mouth

37 7

Total

50

74 14 8

4

1

-7

1

2

Table 11. Overall evaluation of clinical response to each antibiotic regimen

Teicoplanin -t ciprolloxacin ~

..

~~~~~~

(kntaniicin i piprrac.illin ~

.

~~~

Resolution Improved Failed

23 (61%) 5 ( 1 3%) 1 0 (26%))

I 5 (43%) 2 ((1%) 18 (51'x)

Total

38

35

P=O.O5 (Chi-squared) for diffcrencc bctween response rates for two treatment groups.

12

Stephen M. Kelsey et a1

the pathogens in the gentamicin/piperacillin group were eradicated while five (29%) persisted (two Staph. epidermidis, two E. coli, one Staph. aureus; Tables I11 and IV). When the clinical outcome of microbiologically documented infections was considered separately, a favourable response was achieved in 1 7 / 2 3 cases (74%) by teicoplanin plus ciprofloxacin and only 6/17 ( 3 5%) cases with gentami-

Table 111. Microbiological response of each of the 4 3 organisms, isolated from 40 febrile events, to either antibiotic regimen. Erad(icated). Persist(ent) or Ind(eterminant) Teicoplanin + ciprofloxacin

Staph. epidermidis Staph. aureus Sfrep. mitis Sf rep. sanguis Strep. Jaecalis Strep. viridans Enierococcus sp. E. coli Enterobacter cloacae Pseudomonas sp. Aspergillus Jumigatus Candida tropicalis Cytomegalovirus

Gentamicin + piperacillin

Erad

Persist

Ind

Erad Persist

Ind

8

1

3

5

1

2 1

7

1

.1

1 2

1 1 2

1

1

1 1

2 1

1 1

1 1 1

Table IV. Bacteriological evaluation of 40 microbidogically documented episodes

+

Teicoplanin ciprofloxacin Eradicated Persistent Indeterminate

14 (61%) 4 (17%) 5 (22%)

Total

23

Gentamicin + piperacillin

9 ( 5 3%) 5 (29%) 3 (18%) 17

Table V. Clinical response to each antibiotic regimen for microbiologically documented infections only

+

Teicoplanin ciprofloxacin

Gentamicin piperacillin

+

5 (29%)

Resolved Improved Failed

14 (61%) 3 (13%) 6 (26%)

1 (6%) 11 (65%)

Total

23

17

P=0.034 for difference between two groups (Chisquared).

cin plus piperacillin (P=0.034. Chi-squared: Table V). Staph. epidermidis was the primary infecting organism in six of the 11 failures in the gentamicin/piperacillin group (although in only two cases was it documented to have persisted): other isolates were E. coli (three), Strep. miMs and Staph. uureus. In only one of the three cases with primary bacterial infection who failed to respond to teicoplanin plus ciprofloxacin was Staph. epidermidis both the primary pathogen and thought to be the cause of persistent fever. One further patient in this group failed to respond owing to coexistent superinfection with aspergillus and a n enterobacter was isolated in the remaining case. In terms of clinical response patients with Staph. epidermidis infections responded to teicoplanin plus ciprofloxacin in 10 out of 12 cases (83%)whereas only two of eight ( 2 5%) Staph. epidermidis infections responded to gentamicin plus piperacillin (P=0.032). There were three (8%) documented superinfections in patients receiving teicoplanin plus ciprofloxacin (cytomegalovirus pneumonia, Aspergillus pneumonia, Staph. epiderniidis bacteraemia) and two (6%)in patients receiving gentamicin plus piperacillin (two Aspergillus pneumonia). Study drugrelated adverse events occurred in 8/42 patients randomized to receive teicoplanin plus ciprofloxacin ( 19'x). All were elevations of either serum bilirubin or hepatic transaminases and most were no greater than 3 times the upper limit of normal. In no case was it necessary to withdraw study drug therapy. By contrast. treatment-related adverse reactions occurred in 8/38 patients randomized to receive gentamicin plus piperacillin ( 2 I x). Study drug therapy had to be withdrawn in two cases (5%) owing to adverse events: one patient developed acute renal failure and one had an anaphylactic reaction to piperacillin. Three patients developed mild renal impairment and two had small rises in serum transaminases; one patient had a rise in serum creatinine to greater than 200 pmol/l. DISCUSSION Previous studies of febrile neutropenic patients both after chemotherapy and bone marrow transplantation have confirmed the efficacy of teicoplanin against Gram-positive infections (Del Favero et al. 1987: Menichetti et al. 1988). Between 80% and 90% of patients have shown response to first-line antibiotic therapy when teicoplanin was included as part of a n empirical regimen with amikacin and ceftazidime. Preliminary data have shown the combination of teicoplanin plus ciprofloxacin to have an additive anti-staphylococcal effect in vitro and, in vivo. to be an effective empirical regimen for febrile neutropenic patients (C. S. Lewin, unpublished observations). Ciprofloxacin alone has anti-staphylococcal activity despite concerns about the emergence of resistance in coagulase-negative staphylococci with its long-term use (Smith et al, 1988: Oppenheim et al, 1989). We compared the efficacy of teicoplanin plus ciprofloxacin with a standard empirical regimen of gentamicin plus piperacillin. Our data show that the overall response rate for teicoplanin plus ciprofloxacin (74%) was significantly better than that seen with gentamicin plus piperacillin (49%). When microbiologically documented infections were considered separately, however, the response to teicoplanin plus

Randomized Treatment of Fever ciprofloxacin was similar at 74%,and was significantly better than the response to gentamicin plus piperacillin which fell to 3 5%. This is despite the inclusion of three non-bacterial primary infections in the patients randomized to receive teicoplanin. Clinical response, however, did not correlate directly with microbiological evidence of eradication of the primary pathogen. Teicoplanin and ciprofloxacin led to eradication of organisms in 61% of cases while gentamicin plus piperacillin produced eradication in 5 3%: 60% ( 3 / 5 ) persistent organisms in the gentamicin/piperacillin group were Grain-positive whereas these organisms accounted for only 2 ( 1 /4) persisting after therapy with teicoplanin and ciprofloxacin. Neither antibiotic regimen led to a significantly greater superinfection rate than would be expected. Drug-related adverse reactions were also as common for each antibiotic combination. However, only two severe adverse reactions occurred which necessitated withdrawal of study drug therapy, both in patients receiving gentamicin plus piperacillin. In keeping with previous observations (Kelsey rt (11. 1 9 9 0 ) . significant aminoglycoside-associated renal impairment appeared to be unavoidable in a minority of cases despite regular monitoring of serum drug levels. Overall. Gram-positive organisms accounted for 78%, of microbiologically documented infections. with Stnph. r p i d w rnirlis being the predominant pathogen. This represents a 30% increase since 1 9 8 7 in the proportion of Gram-positive organisms isolated as primary pathogens in neutropenic patients in our unit (KelseyPt 01. 1990).As might be expected, the clinical response rate of Stuph. epidPrrnidis infections to teicoplanin plus ciprofloxacin ( 8 3 ' x ) was higher than the surprisingly low figure of 25lX seen for gentamicin plus piperacillin. The poor clinical efficacy of gentamicin and piperacillin against Stoph. cy7iderniidis again raises the question of whether inclusion of a specific anti-Gram-positive agent, such as teicoplanin. in an empirical antibiotic regimen, is justified. The main objections to using vancomycin as a firstline agent are cost. potential nephrotoxicity and the low morbidity from Staph. epidrrmidis infections in which vancoiiiycin is withheld white microbiological data is awaited (Hathorn et d. 1987). Teicoplanin showed no significant toxicity in our study and has been shown previously to be less nephrotoxic than vancomycin (Smith et al. 1 9 8 9 ) . No increased morbidity was seen in our patients with Staph. q~idrrriiidisinfections who failed to respond to gentamicin and piperacillin but this was also true of infections with E. roli. We feel that the high incidence of Gram-positive infections in our patients. togethcr with the low response rate of these infections to gentamicin and piperacillin. probably justifies the first-line use of teicoplanin in our unit. Geographical differences in infection patterns. however, suggest that this policy may not be appropriate for all centres treating patients with haematological malignancy. We conclude that teicoplanin plus ciprofloxacin is more effective than gentamicin plus piperacillin for the empirical treatment of febrile neutropenic patients. This is largely

13

because of the specific anti-Gram-positive action of teicoplanin. This combination provides a n effective, non-toxic regimen for first line therapy, particularly in units where Grampositive organisms predominate.

ACKNOWLEDGMENTS We would like to thank Miss D. Harding and Mrs C. Davis for their assistance in this study. This work was supported by a grant from Merrell Dow Pharmaceuticals Ltd.

KEFEKENCES Buniva. G.. [)el Favero. A,. Bernareggi, A.. Patoia. L. & Palumbo. R. i 1988)Pharmacokineticsof "C-teicoplanin in healthy volunteers. Journal q/ Atitiniicrobial Ch~tnotherapy,21, (Suppl.A ) . 2 3-28. llel Favero. A,. Menichetti. F.. Guerciolini. R.. Bucaneve. G.. Baldelli. F.. Aversa. F.. Terenzi. A,. Davis. S. & Pauluzzi, S. (1987) Prospective randomized clinical trial of teicoplanin for empiric combined antibiotic therapy in febrile. granulocytopenic acute leukaemia patient. Antimicrobial Agents nrrd Chernotherupy. 3 1 , 1 126- 1 1 29. Hathorn. 1.W.. Rubin. R. & Pizzo. P.A. (1987) Empirical antibiotic therapy in febrile neutropenic cancer patients: clinical efficacy and impact of monotherapy. Antimicrobial Agents and Cheniotfierapy. 31, 971-977. M.. Wood. M.E.. Shaw. E.. Jenkins. G.C:& Newland. A.C. i 1990) A comparative study of intravenous ciprofloxacin plus benxylprnicillin versus netilmicin plus piperacillin for the empirical treatment of fever in the neutropenic host. ]octrnn/ oj Antimicrobial C h r t m ~ t l i ~ r q25, y , 149-1 57. Klastersky. 1. ( 1986) Concept of empiric therapy with antibiotic combinations. Anirrii,czn /ournal qf Medicitit,. 80, (5~). 2-1 2. Menichetti. F.. Del Favero. A,. Bucaneve. G.. Aversa. F.. Baldelli. F.. Pelicini, R.. Terenzi. A. & Pauluzzi. S. (1988) Teicoplanin in empirical combined antibiotic therapy of bacteraemia in hone marrow transplant patients. ]onrnnl oJ Antimicrobial Chrtnotlic~rr~pt~. 21, (Suppl. A). 105-112. Oppenheim. B.A., Hartley, 1.W.. Lee. W. & Hurnie. 1.P. (1989) Outbreak of coagulasr negative staphylococci highly resistant to ciprofloxacin in a leukaemia unit. Hritisli Medical /ournu/. 299, 294-297. Schiff, 1.B.. Small, C.J. & Pennington. J.E. (1984) Comparative

activities of ciprofloxacin. ticarcillin and tobramycin against experimental Pseudornorias aeruginosa pneumonia. Atitiniicrobial Agents ond Ch~niotherapy.26, 1-4. Smith. C.M., Leyland. M.J.. Parrell. I.D.& Gedde. A.M. (19x8) A clinical. microbiological and pharmacokineticstudy of ciprofloxacin plus vancomycin as initial therapy of febrile episodes in neutropenic patients. Journal qf Atitiniic.robiul Chemotherapy. 21, 647-65 5 .

Smith. S.K.. Cheesbrough. 1.. Spearing. R. & Davies. J.M. (1989) Kandomized prospective study comparing vancomycin with teicoplanin in the treatment of infections associated with Hickman catheters. Aritimicrobial ADcntsnnd Chrtnoth~rap.y,33, 1 193-1 199.