|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 24, No 25 (September 1), 2006: pp. 4129-4134 © 2006 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.03.9909 Outpatient Oral Antibiotics for Febrile Neutropenic Cancer Patients Using a Score Predictive for Complications
From the Institut Jules Bordet, Brussels, Belgium. Address reprint requests to Marianne Paesmans, MSc, Institut Jules Bordet, Rue Héger-Bordet, 1, B 1000, Bruxelles, Belgium; e-mail: marianne.paesmans{at}bordet.be
Purpose Since febrile neutropenic patients were recognized to constitute a heterogeneous population, several models have been developed for predicting the risk of serious medical complications. The Multinational Association for Supportive Care in Cancer score and its derived clinical prediction rules have been validated, but thus far there were no data about its use for simplifying therapy in predicted low-risk patients. Patients and Methods In a single institution, we followed all episodes of febrile neutropenia between January 1999 and November 2003. Those patients predicted at low risk for complications, who were not receiving antibacterials at fever onset and were eligible for treatment with oral antibiotics, were treated with ciprofloxacin and amoxicillin-clavulanate and were discharged if they were clinically stable or improving after an initial observation period. The primary end point of the study was the rate of resolution of the febrile neutropenic episode without complications, among these early discharged patients. Results Of 383 first febrile neutropenic episodes predicted at low risk of omplication, 178 patients (33 men and 145 women, mainly with solid tumors) were treated orally; they constituted the basis of our analysis. Seventy-nine patients (44%) were discharged early (with a median time to discharge of 26 hours); no complications occurred among them but three patients had to be readmitted, resulting in a success rate of 96% (95% CI, 92% to 100%). Conclusion Our study shows that oral therapy followed by early discharge was feasible in a small but significant proportion of patients selected by a strategy combining predicted low risk and medical and nonmedical criteria.
Febrile neutropenia in cancer patients represents a common complication from chemotherapy that is potentially lethal. Until some years ago, the recommended treatment was empiric broad spectrum intravenous antibiotics and hospitalization for all patients.1,2 However, as febrile neutropenic patients represent a heterogeneous population, among which only a small proportion is at risk of serious complications and death, risk predictive models have been developed and some3,4 have been validated. The goal of these models is to predict risk at the onset of an episode and to identify patients who might be eligible for less intensive antibiotic therapy and/or outpatient management. One model, the Talcott's model, was already used in a pilot study in preparation of a subsequent randomized clinical trial as a clinical tool for guiding management.5 In contrast, the so-called Multinational Association for Supportive Care in Cancer (MASCC) scoring system4 has been validated under various clinical conditions6,7 and has been otherwise widely accepted,8 but has not been demonstrated so far to be useful for selecting patients for oral antibiotic therapy to be given at home (Table 1).
The outpatient setting was tested by different investigators. Malik et al9 compared outpatient with inpatient oral therapy in an unselected population, which included patients with acute leukemia. However, among the outpatients, 21% needed to be readmitted and the mortality rate was 4%. Other studies at The University of Texas M.D. Anderson Cancer Center (Houston, TX)10,11 used both intravenous (IV) and oral regimens in outpatients and reported high response rates and low readmission figures. However, we don't know which proportions of the overall patients with febrile neutropenia were included in these studies. Further, IV treatment can be easily administered at home in the United States, although this is not necessarily true in other countries, adding an additional constraint for developing a simplified therapeutic approach.
Oral antibiotics with early hospital discharge were compared with inpatient intravenous antibiotics in a recently published randomized single-center study.12 The definition of low-risk febrile neutropenia was based on the original definition proposed by Talcott et al,13 but it was felt that more severe exclusion criteria had to be added for predicting low risk. Patients were randomly assigned to two groups: IV antibiotics administered in hospital versus oral antibiotics and discharge from the hospital soon after a mandatory observation period of 24 hours. Rate of effective early discharge was quite high (55% after 2 days) in the oral antibiotics arm. These authors also retrospectively assessed the MASCC score and found that 95% of the patients had scores We performed our investigation in patients with consecutive febrile neutropenic episodes using the MASCC score to predict risk and we tailored management based on predicted risk but not only on this basis.
We screened all febrile neutropenic episodes occurring in cancer patients older than 16 years who were treated with chemotherapy from January 1999 to November 2003. Neutropenia was defined as an absolute neutrophil count of less than 0.5 x 109 cells/L or less than 1.0 x 109 cells/L and expected to fall below 0.5 x 109 within 24 to 48 hours. Fever was defined as a temperature  38.5°C (or 38.0°C twice during a 12-hour interval). All patients were assessed by the MASCC score, which was one of the criteria used to choose treatment.
Patients with a score After giving informed consent, all eligible patients were started on oral antibiotics as soon as the initial work-up, including blood cultures, was completed, and they remained hospitalized for at least 24 hours under close clinical and microbiological surveillance. After this initial observation, patients could be discharged with oral treatment if they were clinically stable or improving and if their home environment and psychosocial status allowed a rigorous follow-up outside of the hospital (phone accessibility and compliance to treatment and to surveillance). In addition, patients' willingness to return home early and the perceived safety of that decision had to be confirmed by the responsible physician. Patients discharged early on oral treatment were instructed to record their temperature once every 6 hours and to come back to the hospital once every 2 days for blood sampling until fever resolution for a period of 5 days. Every other day, the patients were contacted by phone in order to assess for any potential need for readmission. The oral treatment consisted of ciprofloxacin 500 mg tid plus amoxicillin-clavulanate 500 mg tid. As the MASCC score predictive value had already been validated, further validation was not the purpose of this study. Our primary objective was to assess the safety of our procedure for patients who did benefit from early discharge (as it was done by Talcott et al in their pilot study5). Consequently, our primary end point was the resolution rate without any serious medical complications among those patients. Serious medical complications were defined as in our previous paper,4 but for this study we added the need for readmission, as readmitting a patient during the course of a febrile neutropenic episode does not represent a full success of the strategy even in the absence of complications. Secondary objectives included an estimation of the proportion of low-risk patients who could effectively be treated with a simplified oral empiric treatment and an estimation of the proportion of patients who could effectively be discharged on oral treatment; documentation of the reasons for keeping a low-risk patient on oral treatment hospitalized; and patient response to oral treatment, identification of factors predictive of early discharge, and infection documentation in low-risk patients who received an oral treatment. Success to the empiric antibiotic treatment was defined as a return to normal temperature for 5 consecutive days, a clearing of symptoms and signs of infection (if applicable), and pathogen eradication (if applicable). If these criteria were not met, response to treatment was assessed as failure, using a pragmatic intent to treat approach. Infection documentation was classified using European Organisation for Research and Treatment of Cancer International Antimicrobial Therapy Group criteria.14 Our sample size was constructed in order to get a confidence interval for the rate of resolution without serious complications (expected rate around 95%), having a half length lower than 5%. (Here our interval has the form: X – accuracy; X + accuracy. The half length is then the accuracy of the number.) With these conditions, a sample of at least 73 patients who were discharged early was required. We registered all febrile neutropenic episodes in the study but the primary analysis was done on the first episodes per patient, in order to have data that could be considered as independent.
For the analysis, we calculated estimated proportions together with 95% CIs. Comparisons between proportions were made using homogeneity
Overall Patient Characteristics Accrual started in January 1999 and ended in November 2003. During that period, 611 episodes occurred, 441 in patients with a low-risk prediction and 170 in patients with a high-risk prediction, as assessed by the physician who examined the patient in a hospitalization unit or in the emergency room. Of the predicted low-risk episodes, 189 were eligible for oral treatment (43% of episodes; 95% CI, 38% to 48%). The reasons for not administering oral treatment to predicted low-risk patients are described in Table 2.
The 189 episodes with oral treatment prescribed occurred in 178 patients. The 178 (46%) first episodes of a total of 383 first episodes predicted at low risk (95% CI, 41% to 52%) constitute the basis of this report, but to provide an overall picture of our results, Figure 1 illustrates the resolution rates overall and in each considered subgroup. Patients characteristics, MASCC score distribution, and oral treatment administered are listed in Table 3. Due to our selection criteria, almost all patients had a solid tumor, because in our institution, antibacterial prophylaxis is given to most patients with hematologic tumors. Median age of the patients was 53 years, with a majority of female patients with breast tumor, and about half of the patients had no characteristics associated with an unfavorable outcome.
Infection documentation is listed in Table 4. Bacteremia rate (7%) is much lower than in a general population of febrile neutropenic patients and overall rate of microbiological documentation is also low (18%), suggesting that low-risk prediction might be associated to low bacteremia risk.
Early Discharge From Hospital We considered that a patient was discharged early if he stayed in the hospital for fewer than 2 days and had no clinical deterioration during that time. Seventy-nine patients (44%; 95% CI, 37% to 52%) benefited from early discharge and 99 patients (56%) remained hospitalized for various reasons (Table 5). A medical event was documented in 61 patients (62%). The remaining patients stayed hospitalized due to subjective reluctance of the responsible physician, the patient's family, or patient refusal. Overall median time to discharge was 52 hours, whereas the median was 26 hours for the patients with early discharge (first quartile: 20 hours; third quartile: 41 hours) and 137 hours for the patients with episodes requiring prolonged stay in hospital.
Occurrence of Serious Medical Complications For patients treated orally, the rate of resolution without complications was 169/178 (95%; 95% CI, 91% to 98%). Among the 79 patients discharged early, no serious medical complication occurred, as defined in Klastersky et al4; however, three readmissions were required (stomatitis and oesophagitis development requiring a shift to IV antibiotics, persisting fever without therapy change, chills and shift to IV antibiotics). Strategy overall success rate was therefore 76/79 (96%; 95% CI, 92% to 100%). Among the remaining 99 patients, nine developed a serious medical complication, including two patients who died before episode resolution. One death was due to candidemia and the other remained unexplained. In this group of patients, fever resolution without serious complication was 90/99 (91%; 95% CI, 85% to 97%). The success rate was associated to the reason for no discharge; from 79% for patients with persisting fever requiring treatment change, it increased to 91% for patients with another medical reason, and to 96% for patients without medical reason (P = .03). The observed serious medical complications are listed in Table 6.
Response to Oral Treatment Overall, response to oral treatment was 142/178 (80%; 95% CI, 74% to 86%). Response rates for patients with early discharge and for patients remaining hospitalized were 76/79 (96%) and 66/99 (67%), respectively (P < .001).
Are There Factors Predictive for Early Discharge?
Febrile neutropenia affects thousands of patients worldwide. Simplified management including outpatient treatment has great potential economical value and would also contribute to a better quality of life for many patients. There is thus far no indication that validated predictive models3,4 can be used to tailor therapy, safety remaining a key issue.15 Recently, a randomized study12 demonstrated that oral antibiotics in conjunction with early discharge, for patients stable after 24 hours' in-hospital monitoring, offers a feasible and cost-effective alternative to conventional management of low-risk neutropenia. We evaluated this simplified approach in consecutive patients predicted to be at low-risk by the validated MASCC score4 and expected to provide higher sensitivity (for a comparable predictive positive value) in assessing risk. However, we should not amalgamate risk prediction and suitability for oral treatment. Indeed, only 189 (43%) of 441 low-risk patients were eligible for oral therapy as per protocol. Among the reasons for not administering oral amoxicillin-clavulanate plus ciprofloxacin—a well-accepted regimen for that indication14,16—the most frequent was the administration of antibacterial prophylaxis (71%), which largely consisted in a fluoroquinolone administration (ciprofloxacin), as we felt inappropriate to continue therapy with a drug proven unsuccessful as a preventive measure. This led to the exclusion of most patients with hematologic malignancies and constituted a selection independent from our system to predict risk. However, we have to recognize that resolution rate without complication was not as good in predicted low-risk patients treated intravenously as it was in those patients who did receive the oral treatment. Indeed, there were 252 episodes predicted at low risk occurring in 205 patients, for whom treatment was administered intravenously with a resolution rate of 85% without complications, which was significantly different from the rate reached in the population of patients benefiting from oral treatment. This finding probably means that antibacterial prophylaxis or hematologic malignancy, although not taken into account in our predicted risk, is associated with the risk of complications in febrile neutropenia. We should then carefully acknowledge the limitation in the generalizability of our conclusions to institutions in which the same policy for prophylaxis is in use, and we have to further assess our strategy in the patients on antimicrobial prophylaxis but who are otherwise eligible for oral treatment as soon as we have alternatives to the regimen that was administered as oral treatment. We further conclude that we were using a more sensitive validated strategy for predicting low risk than Innes et al,12 but that this selection does not necessarily translate into a higher rate of patients treated orally and discharged early. Unfortunately, we cannot make any comparison, as we do not know the ratio of patients screened (among all febrile neutropenic episodes) to the patients randomly assigned in the Innes study. In contrast, patients with hematologic malignancies are more often hospitalized for their anticancer treatments, and thus often develop febrile neutropenia during hospitalization, making the question of early discharge far less pertinent. Inability to swallow or other contraindications for oral therapy (most often due to mucositis, nausea, and/or vomiting) were represented as 11% and 6%, respectively, of causes for ineligibility of oral treatment, and this further decreases the rate of patients susceptible to be discharged. We based our analysis on the 178 first episodes eligible for oral therapy in order to get independent episodes. Indeed, the MASCC score was developed using a single episode per patient and we are presently investigating whether it remains valid for subsequent episodes. Infection documentation among these 178 patients had a distribution similar to those in recent reports in low-risk patients.14,16 Interestingly, patients with microbiologically documented infection were significantly less likely (P = .05) to be discharged early from the hospital (9 of 32 or 28%) in contrast with 69 (48%) of 145 patients without any identified pathogens. It might be useful in the future to add some parameters for early detection of bacterial infection to the MASCC score, such as CRP or IL8 levels, like in pediatric populations.17 The core question in our study was whether patients with febrile neutropenia identified at low risk and receiving oral empirical antibiotic therapy could be discharged safely from the hospital before episode resolution, while acknowledging that our algorithm makes use of risk assessment and other medical conditions, but also uses nonmedical conditions or ethical issues (such as social criteria and patients' consent to go home) and institution-specific attitudes such as prophylaxis policy. Early discharge feasibility had already been investigated,9,10 but with the use of various clinical criteria for predicting low risk and sometimes with antibiotic regimens given intravenously. Our goal was to further study the question, using a standardized and widely accepted method of risk prediction. Forty-four percent of the patients, who met the criteria for receiving orally administered empirical antibiotic therapy, were discharged early from the hospital. No serious complications occurred among these 79 patients, although three of them had to be readmitted to the hospital. Overall success rate was 96%, confirming that simplified therapy (ie, orally administered empirical antibiotics to ambulatory patients) can be safely used in a small but still worthwhile proportion of patients with febrile neutropenia. Discharge was not mandatory as we knew that we had to face substantial physician anxiety despite the absence of any relevant medical conditions that would prevent a patient from going home. This situation is susceptible to change with the consequence that we will have to go on with the assessment of safety of the procedure. The patients who were not sent back home, despite some of them having achieved the objective criteria for early discharge, constitute a most interesting population. Among these 99 patients, nine patients (9%) developed a serious medical complication and two patients died before febrile neutropenia resolution. Sixty-two percent were not clinically stable or improving, whereas for 38% no objective medical reason was identified justifying prolonged hospitalization, which was attributed to a subjective patient's or physician's belief. Interestingly, the outcome was significantly better when no medical objective justification was available than when a medical reason justified delay for discharge. It should be stressed that our study was conducted in a very pragmatic way, with participation in the study not implying necessarily strong intention from the treating physician to discharge the patient as soon as possible. Finally, we looked at factors associated with early discharge, with only physiological reserve and outpatient status identified as independent factors. We do not know whether the assessment of physiological reserve is reproducible enough, and therefore the information that these covariates provides may lack practical usefulness. Furthermore, inpatient status is probably related to circumstances other than febrile neutropenia, with other reasons to keep a patient in the hospital independent from the predicted risk of complication. Predicting feasibility of early discharge remains then an open issue. The in-hospital observation period is probably very important to selecting those patients to be discharged early as recommended in an Immunocompromised Host Society/MASCC guideline,18 allowing for a significant number of adequate decisions of no discharge, contributing to the success of the strategy. We should also not forget that low-risk prediction and suitability for oral outpatient treatment are two different issues and that we need additional criteria to complement the score. Our purpose was not to show that low-risk prediction means oral outpatient treatment feasibility but rather to use the score as a first step in patients selection. Some additional research is needed to improve upon the rate of patients with effective and safe early discharge.
The authors indicated no potential conflicts of interest.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Hughes WT, Armstrong D, Bodey GP, et al: Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. J Infect Dis 161:381-396, 1990[Medline] 2. Hughes WT, Armstrong D, Bodey GP, et al: Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever: Infectious Diseases Society of America. Clin Infect Dis 25:551-573, 1997[Medline] 3. Talcott JA, Siegel RD, Finberg R, et al: Risk assessment in cancer patients with fever and neutropenia: A prospective, two-center validation of a prediction rule. J Clin Oncol 10:316-322, 1992[Medline] 4. Klastersky J, Paesmans M, Rubenstein EB, et al: The Multinational Association for Supportive Care in Cancer risk index: A multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 18:3038-3051, 2000 5. Talcott JA, Whalen A, Clark J, et al: Home antibiotic-therapy for low-risk cancer-patients with fever and neutropenia: A pilot study of 30 patients based on a validated prediction rule. J Clin Oncol 12:107-114, 1994[Abstract] 6. Uys A, Rapoport BL, Anderson R: Febrile neutropenia: A prospective study to validate the Multinational Association of Supportive Care of Cancer (MASCC) risk-index score. Support Care Cancer 12:555-560, 2004[Medline] 7. Klastersky J: Prospective validation study of the MASCC clinical prediction rule for identification of low-risk patients with febrile neutropenia and for therapeutic decision. Presented at the 6th International Symposium on Febrile Neutropenia, Brussels, Belgium, December 17-19, 2003 (abstr 43) 8. Hughes WT, Armstrong D, Bodey GP, et al: 2002 guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 34:730-751, 2002[CrossRef][Medline] 9. Malik IA, Khan WA, Karim M, et al: Feasibility of outpatient management of fever in cancer-patients with low-risk neutropenia: Results of a prospective randomized trial. Am J Med 98:224-231, 1995[CrossRef][Medline] 10. Rubenstein EB, Rolston K, Benjamin RS, et al: Outpatient treatment of febrile episodes in low-risk neutropenic patients with cancer. Cancer 71:3640-3646, 1993[CrossRef][Medline] 11. Elting LS, Rubenstein EB, Rolston KVI, et al: Outcomes of bacteremia in patients with cancer and neutropenia: Observations from two decades of epidemiological and clinical trials. Clin Infect Dis 25:247-259, 1997[Medline] 12. Innes HE, Smith DB, O'Reilly SM, et al: Oral antibiotics with early hospital discharge compared with in-patient intravenous antibiotics for low-risk febrile neutropenia in patients with cancer: A prospective randomised controlled single centre study. Br J Cancer 89:43-49, 2003[CrossRef][Medline] 13. Talcott JA, Finberg R, Mayer RJ, et al: The medical course of cancer patients with fever and neutropenia: Clinical identification of a low-risk subgroup at presentation. Arch Intern Med 148:2561-2568, 1988 14. Kern WV, Cometta A, De Bock R, et al: Oral versus intravenous empirical antimicrobial therapy for fever in patients with granulocytopenia who are receiving cancer chemotherapy. N Engl J Med 341:312-318, 1999 15. Finberg RW, Talcott JA: Fever and neutropenia: How to use a new treatment strategy. N Engl J Med 341:362-363, 1999 16. Freifeld A, Marchigiani D, Walsh T, et al: A double-blind comparison of empirical oral and intravenous antibiotic therapy for low-risk febrile patients with neutropenia during cancer chemotherapy. N Engl J Med 341:305-311, 1999 17. Santolaya ME, Alvarez AM, Aviles CL, et al: Prospective evaluation of a model of prediction of invasive bacterial infection risk among children with cancer, fever, and neutropenia. Clin Infect Dis 35:678-683, 2002[CrossRef][Medline] 18. Feld R, Paesmans M, Freifeld AG, et al: Methodology for clinical trials involving patients with cancer who have febrile neutropenia: Updated guidelines of the Immunocompromised Host Society/Multinational Association for Supportive Care for Cancer, with emphasis on outpatient studies. Clin Infect Dis 35:1463-1468, 2002[CrossRef][Medline] Submitted August 25, 2005; accepted May 10, 2006.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 tests. Time to discharge was estimated using the Kaplan-Meier method. Logistic regression models were used to identify factors associated with early discharge. All reported P values are two-sided and the threshold for significance was 5%.The protocol was approved by the ethical committee of the institution (Institut Jules Bordet, Brussels, Belgium). 
