Volume 106, Issue 11 p. 2489-2494
Original Article
Free Access

Once daily, oral, outpatient quinolone monotherapy for low-risk cancer patients with fever and neutropenia

A pilot study of 40 patients based on validated risk-prediction rules

Kenneth V. I. Rolston M.D.

Corresponding Author

Kenneth V. I. Rolston M.D.

Division of Internal Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Fax: (713) 745-6839

Section of Infectious Diseases, Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 402, Houston, TX 77030===Search for more papers by this author
Ellen F. Manzullo M.D.

Ellen F. Manzullo M.D.

Division of Cancer Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Linda S. Elting Dr.PH.

Linda S. Elting Dr.PH.

Division of Biostatistics and Applied Mathematics, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Susan E. Frisbee-Hume R.N.

Susan E. Frisbee-Hume R.N.

Division of Internal Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Leslie McMahon R.N.

Leslie McMahon R.N.

Division of Internal Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Richard L. Theriault D.O.

Richard L. Theriault D.O.

Division of Cancer Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Shreyaskumar Patel M.D.

Shreyaskumar Patel M.D.

Division of Cancer Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
Robert S. Benjamin M.D.

Robert S. Benjamin M.D.

Division of Cancer Medicine, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas

Search for more papers by this author
First published: 18 May 2006
Citations: 33

These data were presented in part as abstract 375 at the 41st Annual Meeting of the Infectious Diseases Society of America, San Diego, California, October 9-12, 2003.



The objective of this study was to assess the feasibility of empiric, oral, outpatient quinolone monotherapy in 40 adult patients with fever and neutropenia who were at low risk for serious medical complications.


Patients with breast cancer or sarcoma who presented with fever and neutropenia and were identified as low risk received empiric, oral, quinolone monotherapy (gatifloxacin at a dose of 400 mg once daily). Patients who had a significant source/focus of infection on presentation were excluded. After an initial observation period of 4 to 8 hours in the emergency center, the remainder of their management was ambulatory. Patients were evaluated for response to therapy, development of complications and/or the need for hospital admission, and drug-related adverse events.


Three of 43 patients studied were ineligible medically because of the presence of Common Toxicity Criteria (version 3.0) Grade >2 mucositis. Of the 40 eligible patients, 38 patients (95%) responded to gatifloxacin monotherapy, although 1 patient requested hospital admission (92% response for ambulatory management). The mean duration of therapy was 7 days, and the median number of days from enrollment to defervescence was 4 days. There were no serious medical complications, no drug-related adverse events, and no deaths on study or during 30 days of follow-up.


The results from this study indicated that outpatient quinolone monotherapy in low-risk febrile neutropenic patients is safe, effective, and well received. These conclusions need to be validated in a randomized trial. Cancer 2006. © 2006 American Cancer Society.

The nature and severity of infections that develop in cancer patients with fever and neutropenia require the prompt administration of potent, broad-spectrum, empiric antibiotic therapy.1 Standard practice in the past has been to hospitalize all febrile neutropenic patients until the resolution of fever and neutropenia to facilitate the delivery of supportive care and to ensure close monitoring of patients.2, 3 Although this approach has been effective in reducing infection-related morbidity and mortality, it may lead to unnecessary hospitalization in some patients, exposure to and superinfection with multidrug-resistant nosocomial pathogens in others, and excessive resource utilization and increased costs overall.4

It has been recognized for some time that cancer patients with fever and neutropenia are a very heterogeneous group.5 Two risk-prediction rules recently have been developed and validated.6, 7 Both have a positive predictive value of from 91% to 93% and are being used to identify low-risk neutropenic patients at the onset of febrile episodes.8 This has enabled clinicians to evaluate several alternatives to hospital-based, parenteral therapy, including hospital-based oral therapy,9-11 early discharge on oral or parenteral therapy after initial stabilization in the hospital,12-14 and outpatient parenteral or oral antibiotic therapy for the entire febrile episode.15-18

Most oral, outpatient regimens are quinolone-based (e.g. ciprofloxacin, amoxicillin/clavulanate, or another agent in penicillin-allergic patients), and oral monotherapy with older generation quinolones (ofloxacin, ciprofloxacin) is not recommended.1, 19 The availability of newer generation, broad-spectrum quinolones with enhanced gram-positive activity (e.g. gatifloxacin, moxifloxacin) has made it possible for clinicians to consider monotherapy for low-risk febrile neutropenic patients.20-22 We conducted a pilot feasibility study of oral, outpatient quinolone (gatifloxacin) monotherapy in 40 low-risk cancer patients who presented to the emergency center at The University of Texas, M. D. Anderson Cancer Center (MDACC).


The study was reviewed and approved by The MDACC Clinical Review Committee and the Institutional Review Board. Informed consent was obtained from all patients, and patient confidentiality was maintained in compliance with current Health Insurance Portability and Accessibility Act regulations/guidelines.


All patients were identified and recruited from the Emergency Center of MDACC. This study was designed and implemented prior to the publication of the Multinational Association for Supportive Care in Cancer (MASCC) risk index for the identification of low-risk patients.7 We initially used criteria originally published by Talcott et al. (with some modifications) to define our low-risk patients.12, 16 After publication of the MASCC risk index, MASCC scores also were calculated. Patients in Risk-Group 4 according to the criteria published by Talcott et al. (clinically stable outpatients with no concurrent comorbidity) were enrolled on this trial. We did not exclude patients age ≥ 65 years. We did limit the study to patients who had solid tumors with breast cancer or sarcoma, because these 2 groups accounted for >90% of low-risk patients in our previous studies.16-18 Additional eligibility criteria included ≥1 documented temperature ≥38.3°C and an absolute neutrophil count [ANC] of ≤500/mm3 (or <1000/mm3 with an expected decline to <500/mm3). To further reduce the risk to patients, we added additional medical and nonmedical exclusionary criteria (which also have been part of our previous trials in this setting). Medical criteria included the following: 1) evidence of a significant source/focus of infection (e.g. pneumonia, endocarditis, meningitis, neutropenic enterocolitis, perirectal infection); 2) impaired renal (serum creatinine >2.5 mg/dL and/or creatinine clearance <50 mL/minute) or hepatic (aspirate aminotransferase levels >4 times normal) function; 3) infections known to be caused by mycobacteria, fungi, viruses, protozoal, or bacteria resistant to quinolones at the time of enrollment; and 4) known anaphylactic reaction or hypersensitivity to any fluoroquinolone. We also excluded pregnant or nursing women and patients age <16 years. Nonmedical requirements included the following: 1) residence within a 30-mile radius of MDACC, 2) permission of the patient's primary care physician for enrollment, and 3) ability to cope with protocol requirement, including follow-up clinic visits (family member or care giver at home, 24-hour access to a telephone and transportation, previous history of compliance with treatment protocols).

Treatment Protocol and Clinical Management

All patients were evaluated in our Emergency Center or our Ambulatory Treatment Center. Baseline evaluation included a physical examination, a complete blood count with differential, multichannel biochemical profile, serum electrolyte levels, coagulation profile, urinalysis, and a chest roentgenogram. Blood cultures (1 from the central venous catheters if present and 1 from peripheral blood) were obtained in all patients prior to the institution of antibiotic therapy. Urine, throat, sputum, and cultures from other sites were obtained when clinically indicated. All eligible patients received their first dose of oral gatifloxacin (400 mg) in the Emergency Center or the Ambulatory Treatment Center and were observed for approximately 4 to 8 hours before they were sent home. At the time of discharge after the observation period, patients were given specific instructions (using preprinted instruction sheets) to return to the Emergency Center if they felt worse or if they developed symptoms such as dizziness, confusion, progressive cough and sputum production, decrease in urinary output, or severe headache. They also were provided with a preprinted log to record their temperature 3 times a day, their weight, and a list of other problems. In accordance with the protocol, the patients were seen in various follow-up clinics on Days 2, 3, and 7 of therapy. They were contacted at home by telephone on Days 4, 5, and 6 of therapy by the study coordinator (S.E.F-H.). During these visits and/or conversations, data from the preprinted logs and pertinent clinical and laboratory data were reviewed, and decisions regarding response, failure, drug toxicity, or other adverse events were made. Based on specific criteria, patients either were continued on the treatment protocol (if they were responding to therapy); they were taken off protocol and given alternative therapy, usually in the hospital (if they were not responding); or they were taken off study after the successful completion of therapy.

All patients who had negative cultures and no clinical site(s) of infection (i.e., unexplained fever) were to remain on study for a minimum of 5 days with 4 consecutive afebrile days before the discontinuation of therapy. Patients with negative cultures who remained febrile after 3 days but otherwise were stable clinically were kept on study for 2 additional days. If still were febrile on Day 5 of therapy, then they were considered treatment failures, taken off study, and given alternative therapy based on current clinical practice at our institution. Patients with positive cultures (blood, urine, other sites) were treated until their cultures were rendered negative and they had been afebrile for 4 consecutive days. If cultures remained positive after 3 days of therapy with gatifloxacin, then they were considered treatment failures, removed from study, and treated with appropriate agents based on susceptibility results. Patients with documented central venous catheter-related infections who failed to respond to antibiotic therapy alone after 72 hours had their catheters removed (or exchanged, if removal was not possible because of intravenous access requirements) and remained for 2 additional days on study before they were declared treatment responses or failures. Patients who had negative cultures but clinical sites of infection (e.g. cellulitis, pneumonia) were treated until they had resolution of local infection sites and 4 consecutive afebrile days.

Susceptibility Testing

In vitro susceptibility testing of all isolates from study patients was performed by the institutional microbiology laboratory using previously described methods approved by the National Committee for Clinical Laboratory Standards (NCCLS).23 Gatifloxacin and alternative agents based on preselected, institutional “gram-positive” and “gram-negative” panels were included in the susceptibility testing.

Statistical Analysis

Response rates were described as the proportion of patients who responded and are reported with 95% confidence intervals (95% CI).


Patient Characteristics

Forty-three patients were entered on this study; of these, 3 patients had what were considered protocol violations because of the presence of mucositis Common Toxicity Criteria (version 3.0) Grade >2 mucositis, and were considered ineligible. The characteristics of the 40 evaluable patients are shown in Table 1. There were 21 women and 19 men with a median age of 44 years (range, 20-71 years). The median MASCC score was 24 (range, 16-26).7 Four patients had MASCC scores <21. Thirty-two patients (80%) had an underlying sarcoma, and 8 patients had breast cancer. Most patients (80%) had severe neutropenia at the time of enrollment (ANC <100/mm3). One patient had an ANC of 528/mm3 at enrollment, but the anticipated drop in ANC did not occur, possibly because of the administration of granulocyte–colony-stimulating factor (G-CSF) as part of standard prophylaxis. Overall, 29 patients (72%) were receiving G-CSF at the time of enrollment onto the study.

Table 1. Characteristics of the 40 Evaluable Low-Risk Patients with Fever and Neutropenia
Characteristic* No. of Patients (%)
 Male 19 (47)
 Female 21 (53)
Age, y
 Median 44
 Range 20-71
 Sarcoma 32 (80)
 Breast cancer 8
ANC at randomization
 < 100/mm3 32 (80)
 101-500/mm3 7
Patients receiving G-CSF 29 (73)
  • ANC: absolute neutrophil count; G-CSF: granulocyte–colony-stimulating factor.
  • * Three patients were ineligible because of a protocol violation (> Grade 2 mucositis at the time of enrollment, preventing ingestion of oral therapy).
  • One patient had an ANC of 528/mm3 at the time of enrollment.

Type of Infection and Response

Responses to outpatient, oral gatifloxacin therapy based on types of infection are shown in Table 2. Twenty-three patients (57%) had unexplained fever (i.e., negative cultures and no clinical site of infection). All patients responded to therapy, and none required hospital admission for any reason (i.e., treatment failure, drug toxicity, other complications). Of the 17 patients who had documented infections, 15 patients (88%) responded to therapy. The documented infections included 9 episodes of bacteremia, 4 episodes of upper respiratory tract infection, 2 episodes of urinary tract infection (UTI), and 1 episode each of pneumonia (which developed after enrollment on the study) and cellulitis. The 2 treatment failures included 1 patient with polymicrobial bacteremia caused by Enterococcus faecalis and Flavimonas oxyzihabitans and the patient who had cellulitis (no organism isolated). Both patients responded to alternative antibiotic therapy.

Table 2. Response to Therapy Based on the Nature of the Febrile Episode and Type of Infection
Type of Infection No. of Patients No. of Responses Response Rate (%) 95% CI (%)
Unexplained fever 23 23 100 85-100
Documented infections 17 15 88 64-98
 Bacteremia 9 8 89 52-100
 URTI 4 4 100 40-100
 UTI 2 2 100 16-100
 Pneumonia 1 1 100 2-100
 Cellulitis 1 0 0 0-98
Total 40 38 95 83-99
  • 95% CI: 95% confidence interval; URTI: upper respiratory tract infection; UTI: urinary tract infection.

Eleven patients (27%) had microbiologically documented infections, including 9 episodes of bacteremia (1 polymicrobial), and 2 episodes of UTI. The organisms that were isolated included coagulase-negative Staphylococcus, viridans group streptococci, Micrococcus species, Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae, Pseudomonas aeruginosa, Stenotrophomonas maltophilia, Flavimonas (Pseudomonas) oxyzihabitans, and Enterococcus faecalis (Table 3). Ten patients responded to gatifloxacin therapy (91%; 95% CI, 59-100%). The 1 failure was a patient with polymicrobial bacteremia (E. faecalis and F. oxyzihabitans). He responded to hospital-based, alternative antibiotic therapy (i.e., the addition of vancomycin to cover E. faecalis). One patient with Micrococcus spp. bacteremia responded to gatifloxacin therapy but requested hospital admission after Day 3 because of an inability to cope with the outpatient protocol follow-up schedule. He was considered a response to drug therapy (gatifloxacin) but a failure of outpatient therapy (treatment setting).

Table 3. Response to Therapy in Patients with Microbiologically Documented Infections*
Organism(s) No. of Patients Infection Site Outcome
 CoNS 3 Bloodstream Response
Viridans streptococci 1 Bloodstream Response
 Micrococcus spp. 1 Bloodstream Response
Escherichia coli 2 Bloodstream Response
Escherichia coli 1 Urinary tract Response
 Klebsiella pneumonia 1 Bloodstream Response
Enterobacter cloacae, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia 1 Bloodstream Response
Enterococcus faecalis and Flavimonas oxyzihabitans 1 Bloodstream Failure
  • CoNS: coagulase-negative Staphylococcus.
  • * The overall response rate in patients with microbiologically documented infections (10 of 11 episodes) was 91% (95% confidence interval, 59-100%).
  • This patient responded to gatifloxacin but requested hospital admission because of failure to cope with logistics of outpatient therapy.

Taking all episodes into consideration, 38 of 40 patients (95%; 95% CI, 83-99%) responded to gatifloxacin monotherapy, and 37 of 40 patients (92%; 95% CI, 80-98%) responded to outpatient therapy. All 4 patients who had MASCC scores <21 responded to gatifloxacin, although 1 patient with bacteremia because of Micrococcus species required admission to the hospital.

Other Treatment-Related Issues

All organisms that were isolated, except for P. aeruginosa and E. faecalis from the patients who had polymicrobial bacteremias, were susceptible to gatifloxacin based on NCCLS-designated susceptibility breakpoints. The mean duration of therapy was 7 days (range, 3-14 days). The median time from enrollment to recovery from neutropenia (ANC >500/mm3 for 2 consecutive days) was 4 days (range, 1-18 days). The median number of days from enrollment to defervescence was 3 days (range, 1-6 days). There were no documented bacterial, fungal, or viral superinfections; no admissions to the intensive care unit; no other serious medical complications; no drug-related adverse events; and no deaths on study or within 30 days of the end of therapy (Table 4).

Table 4. Other Treatment-Related Issues in Patients Receiving Oral Gatifloxacin Monotherapy
Issue Result or Outcome
Duration of therapy, days
 Median 7
 Range 3-14
ANC recovery (>500/mm3), days
 Median 4
 Range 1-18
Time to defervescence, days
 Median 3
 Range 1-6
ICU admissions or serious complications None
Drug-related adverse events None
Deaths (within 30 days of EOT) None
  • ANC: absolute neutrophil count; ICU: intensive care unit; EOT: end of therapy.


Previous studies by our group and other investigators have demonstrated that outpatient, oral, antibiotic therapy in carefully selected, low-risk, febrile neutropenic patients is an effective and safe treatment strategy.4, 15-18, 24 The oral regimens studied generally consisted of an early-generation quinolone with limited gram-positive activity (e.g. ofloxacin, ciprofloxacin) in combination with another agent (e.g. amoxicillin/clavulanate or clindamycin) with better gram-positive activity. A recently published metaanalysis of 15 randomized trials of oral versus intravenous antibiotics for the treatment of febrile neutropenic patients concluded that it is safe to offer oral antibiotics to patients who have neutropenia with fever if they are at low risk for mortality.25 Furthermore, although no specific oral regimen was favored, the combination of a quinolone and a second drug that was active against gram-positive bacteria (e.g. amoxicillin/clavulanate) was considered prudent. This approach has been endorsed by the Infectious Diseases Society of America and the National Comprehensive Cancer Network (NCCN), with both organizations cautioning against the use of older generation quinolones as monotherapy in this setting.1, 19

The newer generation quinolones (e.g. gatifloxacin, moxifloxacin) are much more active against most gram-positive organisms than the older generation agents.20-22 They also have excellent coverage against the Enterobacteriaceae and better coverage than ciprofloxacin against nonfermentative, gram-negative bacilli, with the exception of P. aeruginosa.26 A recent review of 757 episodes of neutropenic fever in low-risk patients at our institution documented a very low frequency (<1%) of infections caused by P. aeruginosa in this patient population.27 With this information as background, we considered gatifloxacin an appropriate agent to administer as empiric monotherapy in our low-risk neutropenic patients.

Potential advantages of combination regimens over monotherapy include the following: broad antimicrobial spectrum (including P. aeruginosa in ciprofloxacin-based regimens), the potential for synergy against selected organisms, and the potential for reducing the emergence of resistant organisms. Potential disadvantages of combination regimens include potential for increased toxicity, added costs compared with monotherapy, and reduced compliance because of an increase in the number/frequency of agents used.

The current pilot study has provided information regarding several of the issues listed above. The overall response rates of 95% associated with gatifloxacin monotherapy was at least as high as responses we have achieved with combination oral regimens (88-90%) and similar to the rate (91%) reported from a recently published study that evaluated moxifloxacin in this patient population.28 Although 2 isolates (P. aeruginosa and E. faecalis) from polymicrobial bacteremic infections were resistant to gatifloxacin, both patients responded (1 patient responded to gatifloxacin, and the other patient responded to the addition of specific enterococcal therapy), suggesting adequate, empiric, antimicrobial coverage for low-risk patients. In addition, the frequent use of G-CSF for primary or secondary prophylaxis in these patients may have improved responses because of reduced duration of neutropenia and enhanced neutrophil function. Overall, the regimen was tolerated relatively well with no documented antibiotic-related adverse events. This is in contrast to several studies that used combination regimens, in which adverse events (particularly gastrointestinal events) were reported to occur in approximately 29% to 33% of patients, leading to the discontinuation of therapy in a substantial number of patients.10, 11 These events most probably were related to the amoxicillin/clavulanate component of the combinations and were not observed in our pilot study of quinolone monotherapy.

Approximately 14% to 18% of low-risk patients who present at our emergency center with fever and neutropenia are allergic to penicillins or other β-lactam agents, which renders them ineligible to receive the standard combination of ciprofloxacin and amoxicillin/clavulanate (unpublished data from MDACC pharmacy log book). Quinolone monotherapy, if it is determined to be safe and effective in larger randomized trials, may provide a practical therapeutic option for such patients.

The longer half-life of newer generation quinolones, such as gatifloxacin and moxifloxacin, makes once-a-day administration possible. Whether this is an advantage over other regimens that need to be given 2 or 3 times daily by increasing compliance also will have to be evaluated in prospective, randomized trials. However, if once-daily gatifloxacin and multidrug regimens prove to have similar effectiveness in comparative trials, then they may be associated with both increased compliance and some cost savings.

The results of the current pilot study are promising. However, no pilot study can fully assess a novel therapeutic approach. This can be done only by conducting a large, multicenter, randomized, clinical trial comparing standard therapy with the novel approach using internationally accepted clinical trials methodology.5 Based on our pilot study, we believe that a trial comparing standard outpatient oral combination therapy with quinolone monotherapy would be appropriate and timely in our changing healthcare environment.