This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Klaassen, R. J. Articles by Doyle, J. J. Search for Related Content PubMed PubMed Citation Articles by Klaassen, R. J. Articles by Doyle, J. J. Social Bookmarking                            What's this?

"Low-Risk" Prediction Rule for Pediatric Oncology Patients Presenting With Fever and Neutropenia

From the Department of Pediatrics, Children’s Hospital of Eastern Ontario and Children’s Hospital of Eastern Ontario Research Institute, University of Ottawa, Ottawa, Ontario, and Department of Radiology and Division of Hematology/Oncology, Department of Pediatrics, The Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada.

Address reprint requests to Robert Klaassen, MD, Department of Paediatrics, Children’s Hospital of Eastern Ontario, 401 Smyth, Ottawa, Ontario, K1H 8L1 Canada; email rklaassen{at}cheo.on.ca

	ABSTRACT

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To prospectively derive and validate a clinical prediction rule to allow a more tailored approach to the management of pediatric oncology outpatients presenting with fever and neutropenia.

PATIENTS AND METHODS: The clinical prediction rule was derived over a 1-year period and then validated over the following 8 months in a new set of fever and neutropenia episodes. Patients were excluded if they presented with comorbidity or an abnormal chest x-ray (CXR).

RESULTS: Significant bacterial infection (SBI; defined as any blood or urine culture positive for bacteria, interstitial or lobar consolidation on CXR, or unexpected death from infection) was documented in 43 of the 227 episodes. Multivariate analysis found four significant factors: bone marrow disease, general appearance unwell on initial examination, monocyte count less than 0.1 x 10⁹/L, and peak oral or oral equivalent temperature greater than 39°C. Only the monocyte count contributed to determining a low-risk group, excluding SBI with 84% sensitivity (95% confidence interval [CI], 61% to 100%), 42% specificity (95% CI, 38% to 46%), and a negative predictive value of 92% (95% CI, 76% to 100%). If the monocyte count was 0.1 x 10⁹/L at the time of presentation (low risk), the incidences of SBI and bacteremia were 8% and 5%, respectively, versus 25% and 17% in the high-risk group. When validated in a new population of 136 episodes of fever and neutropenia, the incidences of SBI and bacteremia in the low-risk group were 12% and 5%, respectively, and 25% and 19% in the high-risk group.

CONCLUSION: Pediatric oncology outpatients with fever and neutropenia who present with an initial monocyte count of 0.1 x 10⁹/L and do not have comorbidity or an abnormal CXR at the time of presentation are at lower risk for SBI and can be considered for less aggressive initial therapy.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
FEVER WITH NEUTROPENIA is one of the most common admitting diagnoses for pediatric oncology patients, second only to admissions for chemotherapy (admission data, Hospital for Sick Children [HSC], Toronto, Canada). Standard therapy for cancer patients with fever and an absolute neutrophil count (ANC) less than 0.5 x 10⁹/L (500/mm³) has traditionally been hospital admission for administration of broad-spectrum intravenous (IV) antibiotics until the patient is afebrile and the ANC has recovered.¹ This approach has resulted in an infection-related mortality rate of 4% to 6% in adults and 0.4% to 1% in children.^2,3 Although the low pediatric mortality rate is admirable, for many children this results in frequent and lengthy inpatient stays. One report that observed patients over a 2.5-year period noted a mean 3.2 admissions per patient, each lasting for 13 days on average.⁴ The marked increase in multidrug-resistant organisms over the past 10 years from the indiscriminate use of broad-spectrum antibiotics is an additional concern.^5-7

When a patient is neutropenic with a fever, the major concern remains occult bacterial infection, with most series reporting an incidence of bacteremia between 10% and 24%.^4,8,9 A few centers have begun selecting patients at low risk for bacteremia for initial outpatient management, using criteria derived from a combination of personal experience and previously published research.^10-12 The studies on which these criteria were based had looked for and identified different sets of significant variables, with many of these variables not available at the time of presentation.^4,8,13,14 Oncologists are rightly hesitant about adopting this outpatient approach without strong evidence that a low-risk group can be clearly defined.

In the only study that has prospectively validated their "low-risk" criteria, Talcott et al,¹⁵ in 1992, found that only 5% of adult outpatients with cancer in remission and no comorbidity had a serious medical complication, as compared with 34% of patients who did not fit these criteria. Because a 5% prevalence of serious medical complications is considered by many physicians unacceptable to undertake outpatient therapy, we elected to further refine these criteria in a pediatric population. Using the criteria generated by Talcott et al as our starting point, we set out to prospectively derive and validate a clinical prediction rule that would identify a group of pediatric oncology patients presenting with fever and neutropenia who were at low risk for significant bacterial infection.

	PATIENTS AND METHODS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Derivative Set
All pediatric oncology patients consecutively admitted to the HSC from August 1, 1996, to July 31, 1997, with the diagnosis of fever and neutropenia or a specific infection were assessed for eligibility. Fever was defined as an oral or equivalent temperature greater than 38.5°C once, or greater than 38°C on two or more occasions, during a 12-hour period. Neutropenia was defined as an ANC less than 0.5 x 10⁹/L or between 0.5 and 1.0 x 10⁹/L and expected to fall.^16,17 HSC is a tertiary- and quaternary-care pediatric teaching institution that also provides primary oncology care to the metropolitan Toronto area, a population of approximately 5 million. After interviewing the primary care oncology nurses, we estimate that only 10% to 15% of patients from the metropolitan area are admitted to a hospital outside of HSC. Before recommending admission to a hospital outside of HSC, the physician assessing the patient must deem him or her to be clinically stable. Any patients who subsequently become unstable are immediately transferred to HSC.

Potential patients were detected using the daily oncology admission list, which is sent to the office of the hematology/oncology data managers the morning after admission. Any patient admitted with the diagnosis of fever, neutropenia, or infection, such as cellulitis, pneumonia, or sepsis, was assessed. Patients were excluded if they presented with a new diagnosis of cancer or had undergone bone marrow or stem-cell transplantation within the preceding 6 months. All oncology admissions to the intensive care unit were specifically reviewed to ensure that no significant infections were missed.

Predictive Factors
A systematic review was conducted to identify risk factors for significant bacterial infection. Using the OVID search engine on MEDLINE, we performed a literature search of the period from 1986 to 1996, using the following Medical Subject Heading terms: "neutropenia" with "fever" or "infection." A hand search was performed on the reference lists of relevant articles. Expert opinion from three pediatric oncologists and one pediatric infectious diseases specialist was used for additional items and for choosing between comparable items generated from the literature.

Exclusion criteria included inpatient status and comorbidity at the time of presentation, because they had been previously validated as important high-risk factors.¹⁵ Patients with interstitial infiltrate or lobar consolidation on chest x-ray (CXR) at the time of presentation were also excluded because this is one of the outcome criteria.¹⁸ Comorbidity is defined, as per Talcott et al, as another medical condition that independently requires inpatient observation. Predefined conditions meeting these criteria included hypotension requiring fluid resuscitation, hypoxia requiring supplemental oxygen, respiratory distress, pain necessitating IV narcotics, and vomiting or diarrhea requiring fluid resuscitation. Because of differences between pediatric and adult malignancies, we did not exclude patients whose cancer was not in remission but analyzed it as a predictive variable.

The principal investigator (R.J.K.), who collected prediction variables shortly after admission, was blind to outcome status. Information was obtained from the medical record, from laboratory computer database, or by patient or parent interview, if necessary. Thirteen variables were collected on each patient: age,³ presence of bone marrow disease,¹⁴ central venous catheter type,³ general appearance on initial examination, previous granulocyte colony-stimulating factor (G-CSF) therapy,¹⁹ initial ANC, initial lymphocyte count,²⁰ initial monocyte count,⁸ initial platelet count,¹¹ presence of localized bacterial infection on initial examination, peak temperature,^3,8 tumor type,^2,4,21 and sex. Expectation of prolonged neutropenia (> 7 days)²² was not included as a variable because it was considered to be unreliable. Time since chemotherapy was not included because patients with acute lymphocytic leukemia (ALL), the most common tumor in our population, receive daily chemotherapy.

Variables were defined as follows. Bone marrow disease was considered to be present if greater than 5% blast cells or metastatic solid tumor was seen in the last bone marrow study done before admission. Central venous catheter type was categorized as either none, implanted reservoir, or external line. General appearance on initial examination was classified as either well or unwell (septic patients were considered to have comorbidity and were therefore excluded). If in the initial examination the admitting physician made no specific comment about the patient’s general appearance, or if the only comment was "tired" or "pale," the patient was classified as well. If no specific comment was made and negative statements were noted, such as "irritable" or "lethargic," the patient was classified as unwell. G-CSF therapy comprised administration of the drug during the period before presentation with fever and neutropenia.

Initial blood results were determined by 100-cell manual differential counts, with the technologist blind to patient outcome. If the WBC count was 0.3 to 0.5 x 10⁹/L, the differential count was determined by the average of a 50-cell manual differential count carried out independently by two investigators (R.J.K. and J.J.D.) blind to outcome. If the WBC count was 0.2 x 10⁹/L, a differential count was not performed and the ANC, monocyte count, and lymphocyte count were assigned a value of zero. Localized bacterial infection on initial examination was defined as any suspected bacterial focus of infection that required a full course of antibiotic therapy. Peak temperature was the highest value of the home and emergency department/clinic temperature. Patients were given acetaminophen 10 to 15 mg/kg/dose if the temperature exceeded 38.5°C. As is the practice at HSC, axillary temperatures were converted to their oral equivalent by adding 1.0°C.

Outcome
Significant bacterial infection was considered to be the most appropriate outcome measure because the purpose of this clinical prediction rule is to select patients who can be treated safely as outpatients. Significant infection was defined as any blood or urine culture positive for bacteria, interstitial or lobar consolidation on CXR, or unexpected death from infection (patient not palliative) before ANC recovery (> 0.5 x 10⁹/L). A urine culture was considered positive for bacteria if more than 10 x 10⁶ organisms/L were seen independent of the urine WBC count or if more than 1 x 10⁶ organisms/L were seen with 10 WBCs/mm³. Bacteruria was considered a significant infection because outpatient treatment of a neutropenic patient with a urinary tract infection would not be appropriate due to the risk of sepsis. Cultures were determined to be positive by an automated continuous monitoring system (BACTEC 9240; Becton Dickson, Sparks, MD), and no attempt was made to interpret if the specimens might be contaminated. All CXRs were reviewed by one radiologist (T.R.G.) and interpreted with the radiologist blind to the predictive variables. Culture and radiology results were closely monitored in all patients to be certain that other important infections not part of the original definition were not missed.

Reliability
During the study period, two investigators (R.J.K. and J.J.D.) independently assigned predictive variables in 32 consecutive patients to assess reliability. Data collected included comorbidity, general appearance on initial examination, localized bacterial infection on initial examination, peak temperature, and the monocyte count when the WBC count was 0.2 to 0.5 x 10⁹/L. In addition, expectation of prolonged neutropenia (> 7 days) was evaluated because it has been used in previous studies as a predictive factor.²²

The reliability of the CXR results was calculated by comparing the interpretation of an independent reviewer (T.R.G.) to the official radiology report. The reliability of the manual monocyte count was only done when the WBC count was low, since it is well established in patients with a normal WBC count, with one study reporting a correlation coefficient of 0.758.²³ Data were not collected on age, sex, tumor type, bone marrow disease, central line type, or G-CSF therapy, because they were considered to be reliable.

Validation Set
From August 1, 1997, to April 30, 1998, the prediction rule was prospectively validated on a new set of fever and neutropenia episodes. All pediatric oncology patients consecutively admitted to HSC with the diagnosis of fever and neutropenia were assessed using the same exclusion criteria as were used in the derivation set. The variables and outcome were collected blind, using the methods described in the derivation set.

Management of Fever and Neutropenia
All patients were seen in the oncology clinic or emergency department and had a history taken and underwent a physical examination. The reviewing physician was not specified and varied from a junior resident to a staff oncologist. A complete blood count and manual differential count were performed. A peripheral-blood culture and a blood culture from all central-line lumens were obtained. Other studies, such as cultures of the urine, throat, stool, nasopharynx, and central catheter exit site or chest radiology, were performed at the discretion of the admitting physician. All patients were admitted and treated initially with broad-spectrum antibiotics (piperacillin 200 mg/kg/d divided every 6 hours and gentamicin 7.5 mg/kg/d divided every 8 hours, or a similar combination). Other antibiotics may have been added if there was suspicion of a localized infection. Patients were discharged home when either the ANC was greater than 0.5 x 10⁹/L or they met the following criteria: afebrile for longer than 24 hours, negative blood cultures, absence of clinical sepsis, and underlying cancer in bone marrow remission.

Statistical Analysis
We first summarized patient characteristics for both the derivative and validation populations. Next, using the derivative set, we compared the groups with and without significant bacterial infection with respect to patient characteristics. All continuous variables except age were converted to binary variables using predefined thresholds taken from the literature or recursive partitioning. The thresholds used included ANC greater than 1.0 x 10⁹/L,⁴ lymphocyte count greater than 0.7 x 10⁹/L,²⁰ and platelet count greater than 75 x 10⁹/L.¹¹ The thresholds for monocyte count and peak temperature were determined using recursive partitioning, with the resulting number rounded to the nearest clinically appropriate threshold. Tumor type was converted to acute myelogenous leukemia (AML) and non-Hodgkin’s lymphoma (NHL) versus other types, because these patient groups receive highly aggressive chemotherapy.

Possible difference between groups on each univariate variable was checked using Student’s t test or the ² test, as appropriate. Factors significant at a level of .1 were included in the prediction rule development. This exercise consisted of two steps. First we quantified the potential risk associated with each candidate factor using logistic regression (ie, odds ratios and 95% confidence intervals [CIs]). Next, recursive partitioning was used to develop the risk stratification, because it is better suited to develop a rule with high sensitivity.²⁴ To capture the correlation structure between successive episodes from an individual patient,²⁵ we performed a logistic regression with random effect using an SAS macro (GLIMMIX).²⁶ Interobserver reliability was assessed using the unweighted kappa coefficient for binary variables and Spearman’s rho for continuous variables.

	RESULTS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Deriving the Prediction Rule
From August 1, 1996, to July 31, 1997, 156 pediatric oncology patients were admitted with 270 episodes of fever and neutropenia. Twenty-four episodes (9%) were excluded because of comorbidity (severe mucositis requiring IV narcotics, n = 7; severe abdominal or rectal pain, n = 4; respiratory distress, n = 4; dehydration, n = 3; and other causes, n = 6). An additional 19 episodes (7%) were excluded because of an abnormal CXR at the time of presentation (localized consolidation, n =15; interstitial infiltrate, n = 4). One of the excluded patients was palliative and died of congestive heart failure before neutrophil recovery.

Two hundred twenty-seven episodes of fever and neutropenia in 140 patients were used for analysis. Characteristics are outlined in Table 1. Forty-three significant bacterial infections (19%) occurred before neutrophil recovery (28 patients had positive blood cultures [12%], 11 had positive urine cultures [5%], and seven had positive CXRs [3%]). One patient died during the induction phase of standard-risk ALL treatment, 11 days after admission for fever and neutropenia. On autopsy, he was found to have disseminated Pseudomonas aeruginosa infection, despite repeated negative blood cultures and receiving appropriate broad-spectrum, antipseudomonal antibiotics.

View larger version (28K): [in this window] [in a new window]   Fig 1. Tree diagram generated from the derivation set using recursive partitioning. Variables were selected to provide the widest split between branches. One patient did not have an initial monocyte count. SBI, significant bacterial infection; Mono, monocyte count (x 109/L).

Validating the Prediction Rule
From August 1, 1997, to April 30, 1998, 102 pediatric oncology patients were admitted with 161 episodes of fever and neutropenia. Twenty-three episodes (14%) were excluded because of comorbidity, and two episodes (1%) were excluded because of an abnormal CXR at the time of presentation. The remaining 136 episodes were used for validation of the prediction rule (Table 1). Twenty-seven episodes (20%) resulted in significant bacterial infection, including a positive blood culture in 19 episodes (14%), a positive urine culture in three episodes (2%), and a positive CXR in five episodes (4%). One patient with relapsed AML died 3 months after admission without having documented recovery of the ANC.

Using the previously derived risk stratification, the incidence of significant bacterial infection/bacteremia was 12%/5% if the patient was low risk (monocyte count 0.1 x 10⁹/L) versus 25%/22% if the patient was high risk (monocyte count < 0.1 x 10⁹/L). If the initial monocyte count was 0.1 x 10⁹/L, significant bacterial infection was excluded with 74% sensitivity (95% CI, 49% to 100%), 46% specificity (95% CI, 41% to 52%), and a negative predictive value of 88% (95% CI, 71% to 100%) (Table 5). The temperature variable did not improve the sensitivity or the specificity in the validation group (Fig 2).

View larger version (31K): [in this window] [in a new window]   Fig 2. Diagram shows the significant bacterial infection and bacteremia in the validation set according to risk stratification generated from the derivative set. SBI, significant bacterial infection; Mono, monocyte count (x 109/L).

	DISCUSSION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

Other large studies in pediatric oncology patients, most notably the studies by Lucas et al⁴ and Hann et al,³ were unable to develop a prediction model with adequate discriminant ability. Neither study reported looking at the initial monocyte count. The initial ANC turned out to be a poor predictor of bacterial infection, which corroborates the findings of Rackoff et al⁸ and Jones et al.²¹ We found that expectation of prolonged neutropenia (> 7 days), which has been used by Malik et al^22,27 as an exclusion criterion, is very unreliable (kappa = 0.05) and should be not be used in future studies of outpatient therapy.

The results of this study are consistent with the findings of Rackoff et al,⁸ who also noted the predictive power of the monocyte count. In a prospective study of 115 admissions for fever and neutropenia at the James Whitcomb Riley Hospital for Children in Indiana, patients with a monocyte count 0.1 x 10⁹/L had a low incidence of bacteremia (0%). If the monocyte count was less than 0.1 x 10⁹/L and the temperature was less than 39°C, the incidence of bacteremia was 19%; if the temperature was 39°C, it was 48%. Rackoff et al validated the model on a retrospectively collected population and found 0% bacteremia in the high-monocyte-count group, 4% bacteremia in the intermediate-risk group, and 43% bacteremia in the high-risk group. It is reassuring that, using recursive partitioning, we identified similar thresholds for both the monocyte count and temperature. We were unable to show that high temperature was important as a high-risk variable in our validation set, because patients who presented with a low monocyte count and a temperature greater than 39°C had essentially the same incidence of significant bacterial infections as those with a low temperature (26% and 24%, respectively).

A group in Dallas has consistently shown that a sustained increase in the WBC count, ANC, absolute phagocyte count (ANC + monocyte count), and/or platelet count is an early indicator of bone marrow recovery and that patients with rising counts are at low risk for recurrent fever and bacteremia.^13,28 A rising monocyte count has been well described as a predictor of imminent recovery from neutropenia.²⁹ Most patients do not have regular differential counts before presentation with fever and neutropenia. Therefore, it is impractical to use increasing blood counts in the clinic or emergency department as an indicator. It is likely that the monocyte count acts as a surrogate marker of early bone marrow recovery.

The primary concern for oncologists remains the rare but real possibility of death from sepsis. Three patients died before ANC recovery during the 18 months of study. During that period, a total of 431 admissions for fever and neutropenia occurred at HSC, resulting in a mortality rate of 0.7%. One patient was palliative, another died 3 months after admission with relapsed AML, and a third patient died 11 days after admission during the induction phase of standard-risk ALL treatment. In this last patient, disseminated Pseudomonas aeruginosa infection was found on autopsy, despite repeated negative cultures and receiving appropriate antipseudomonalantibiotics. All three patients would have been excluded or classified as high risk at the time of initial presentation.

One limitation of this study is the method of data collection. Because of practical constraints, the data were collected from patients’ charts after admission. Two of the variables, general appearance on initial examination and localized bacterial focus, depended on the written report of the emergency department/clinic physician. We were able to reliably extract the general appearance on initial examination from the chart (kappa = 0.69), but whether this is equally reliable in clinical practice remains to be determined. Fortunately, the monocyte count is unaffected by this method of data collection, and we have clearly demonstrated that it is reliable even when the WBC count is low.

The results of this study will allow clinicians to determine a low-risk subset of pediatric patients with fever and neutropenia at the time of presentation to the clinic or emergency department. This subset of patients could then be considered for less intensive therapy. We are presently conducting a pilot study of once-daily IV antibiotics administered through the clinic until the culture results become available, to determine the feasibility of outpatient therapy in this low-risk group of patients.

	REFERENCES

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
1. Pizzo PA, Robichaud KJ, Gill FA, et al: Duration of empiric antibiotic therapy in granulocytopenic patients with cancer. Am J Med 67:194-200, 1979[Medline]

2. Leong DC, Kinlay S, Ackland S, et al: Low-risk febrile neutropenia in a medical oncology unit. Aust N Z J Med 27:403-407, 1997[Medline]

3. Hann I, Viscoli C, Paesmans M, et al: A comparison of outcome from febrile neutropenic episodes in children compared with adults: Results from four EORTC studies—International Antimicrobial Therapy Cooperative Group (IATCG) of the European Organization for Research and Treatment of Cancer (EORTC). Br J Haematol 99:580-588, 1997[Medline]

4. Lucas KG, Brown AE, Armstrong D, et al: The identification of febrile, neutropenic children with neoplastic disease at low risk for bacteremia and complications of sepsis. Cancer 77:791-798, 1996[Medline]

5. From the Centers for Disease Control and Prevention: Nosocomial enterococci resistant to vancomycin–United States, 1989-1993. JAMA 270:1796, 1993[Free Full Text]

6. Recommendations for preventing the spread of vancomycin resistance: Recommendations of the Hospital Infection Control Practices Advisory Committee (HICPAC). MMWR Morb Mortal Wkly Rep 44:1-13, 1995

7. Lam S, Singer C, Tucci V, et al: The challenge of vancomycin-resistant enterococci: A clinical and epidemiological study. Am J Infect Control 23:170-180, 1995[Medline]

8. Rackoff WR, Gonin R, Robinson C, et al: Predicting the risk of bacteremia in children with fever and neutropenia. Oncol 14:919-924, 1996

9. Doyle JJ, King SM, Comay SA, et al: Oral antibiotic therapy for "low risk" febrile neutropenic episodes (FNE). Pediatr Res Prog 39:154A, 1996 (abstr)

10. Malik IA, Abbas Z, Karim M: Randomised comparison of oral ofloxacin alone with combination of parenteral antibiotics in neutropenic febrile patients. Lancet 339:1092-1096, 1992 (published erratum appears in Lancet 340:128, 1992)[Medline]

11. Mustafa MM, Aquino VM, Pappo A, et al: A pilot study of outpatient management of febrile neutropenic children with cancer at low risk of bacteremia. J Pediatr 128:847-849, 1996[Medline]

12. Preis S, Gobel U, Jurgens H: Outpatient treatment with ceftriaxone alone or in combination with teicoplanin in febrile neutropenic children and adolescents with cancer. J Pediatr 130:500-501, 1997[Medline]

13. Griffin TC, Buchanan GR: Hematologic predictors of bone marrow recovery in neutropenic patients hospitalized for fever: Implications for discontinuation of antibiotics and early discharge from the hospital. Pediatr 121:28-33, 1992

14. Jones GR, Konsler GK, Dunaway RP, et al: Risk factors for recurrent fever after the discontinuation of empiric antibiotic therapy for fever and neutropenia in pediatric patients with a malignancy or hematologic condition. J Pediatr 124:703-708, 1994[Medline]

15. 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. Oncol 10:316-322, 1992

16. From the Immunocompromised Host Society: The design, analysis, and reporting of clinical trials on the empirical antibiotic management of the neutropenic patient—Report of a consensus panel. J Infect Dis 161:397-401, 1990[Medline]

17. Viscoli C, Bruzzi P, Glauser M: An approach to the design and implementation of clinical trials of empirical antibiotic therapy in febrile and neutropenic cancer patients. Cancer 31A:2013-2022, 1995

18. Wasson JH, Sox HC, Neff RK, et al: Clinical prediction rules: Applications and methodological standards. N Engl J Med 313:793-799, 1985[Abstract]

19. Pui CH, Boyett JM, Hughes WT, et al: Human granulocyte colony-stimulating factor after induction chemotherapy in children with acute lymphoblastic leukemia. N Engl J Med 336:1781-1787, 1997[Abstract/Free Full Text]

20. Blay JY, Chauvin F, Le Cesne A, et al: Early lymphopenia after cytotoxic chemotherapy as a risk factor for febrile neutropenia. J Clin Oncol 14:636-643, 1996[Abstract/Free Full Text]

21. Jones GR, Konsler GK, Dunaway RP, et al: Infection risk factors in febrile, neutropenic children and adolescents. Hematol Oncol 13:217-229, 1996

22. 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[Medline]

23. Goossens W, Van Hove L, Verwilghen RL: Monocyte counting: Discrepancies in results obtained with different automated instruments. J Clin Pathol 44:224-227, 1991[Abstract/Free Full Text]

24. Laupacis A, Sekar N, Stiell IG: Clinical prediction rules: A review and suggested modifications of methodological standards. JAMA 277:488-494, 1997[Abstract/Free Full Text]

25. Diggle PJ, Liang KY, Zeger SL: Analysis of Longitudinal Data. New York, NY,Oxford University Press, 1994

26. SAS Institute, Inc:SAS System for Mixed Modules. Cary, NC,SAS Institute, Inc, 1996, pp 423-460

27. Malik IA, Khan WA, Aziz Z, et al: Self-administered antibiotic therapy for chemotherapy-induced, low-risk febrile neutropenia in patients with nonhematologic neoplasms. Clin Infect Dis 19:522-527, 1994[Medline]

28. Aquino VM, Tkaczewski I, Buchanan GR: Early discharge of low-risk febrile neutropenic children and adolescents with cancer. Clin Infect Dis 25:74-78, 1997[Medline]

29. Mullen CA, Buchanan GR: Early hospital discharge of children with cancer treated for fever and neutropenia: Identification and management of the low-risk patient. J Clin Oncol 8:1998-2004, 1990[Abstract]

Submitted May 12, 1999; accepted August 25, 1999.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				R Phillips, R Skinner, and J C Chisholm Treating low-risk febrile neutropenia: Jenny's story Arch. Dis. Child., January 1, 2007; 92(1): 7 - 8. [Full Text] [PDF]

				S. K. Basu, I. D. Fernandez, S. G. Fisher, B. L. Asselin, and G. H. Lyman Length of Stay and Mortality Associated With Febrile Neutropenia Among Children With Cancer J. Clin. Oncol., November 1, 2005; 23(31): 7958 - 7966. [Abstract] [Full Text] [PDF]

				G. H. Lyman, C. H. Lyman, O. Agboola, and for the Anc Study Group Risk Models for Predicting Chemotherapy-Induced Neutropenia Oncologist, June 1, 2005; 10(6): 427 - 437. [Abstract] [Full Text] [PDF]

				M. E. Santolaya, A. M. Alvarez, C. L. Aviles, A. Becker, J. Cofre, M. A. Cumsille, M. L. O'Ryan, E. Paya, C. Salgado, P. Silva, et al. Early Hospital Discharge Followed by Outpatient Management Versus Continued Hospitalization of Children With Cancer, Fever, and Neutropenia at Low Risk for Invasive Bacterial Infection J. Clin. Oncol., September 15, 2004; 22(18): 3784 - 3789. [Abstract] [Full Text] [PDF]

				L. Vidal, M. Paul, I. Ben dor, K. Soares-Weiser, and L. Leibovici Oral versus intravenous antibiotic treatment for febrile neutropenia in cancer patients: a systematic review and meta-analysis of randomized trials J. Antimicrob. Chemother., July 1, 2004; 54(1): 29 - 37. [Abstract] [Full Text] [PDF]

				A. Gassas, R. Grant, S. Richardson, L. L. Dupuis, J. Doyle, U. Allen, O. Abla, and L. Sung Predictors of Viridans Streptococcal Shock Syndrome in Bacteremic Children With Cancer and Stem-Cell Transplant Recipients J. Clin. Oncol., April 1, 2004; 22(7): 1222 - 1227. [Abstract] [Full Text] [PDF]

				W. Rackoff, P. Breitfeld, and M. E. Santolaya Risk Factors in Children With Fever and Neutropenia J. Clin. Oncol., November 15, 2001; 19(22): 4270 - 4271. [Full Text]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Klaassen, R. J. Articles by Doyle, J. J. Search for Related Content PubMed PubMed Citation Articles by Klaassen, R. J. Articles by Doyle, J. J. Social Bookmarking                            What's this?