|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2001 American Society for Clinical Oncology Prospective, Multicenter Evaluation of Risk Factors Associated With Invasive Bacterial Infection in Children With Cancer, Neutropenia, and FeverByFrom the Department of Pediatrics, Hospital Luis Calvo Mackenna; Department of Pediatrics, Hospital San Juan de Dios; Department of Pediatrics, Hospital Sótero del Río; Department of Pediatrics, Hospital Roberto del Río; Microbiology Program, Faculty of Medicine, Universidad de Chile; Department of Pediatrics, Hospital Exequiel González Cortés; and Subcommittee of Infectious Diseases of the National Child Program of Antineoplastic Drugs, Santiago, Chile. Address reprint requests to María Elena Santolaya, MD, Los Huasos 1948, Las Condes, Santiago, Chile; email: msantola{at}machi.med .uchile.cl.
PURPOSE: To identify clinical and laboratory parameters present at the time of a first evaluation that could help predict which children with cancer, fever, and neutropenia were at high risk or low risk for an invasive bacterial infection. PATIENTS AND METHODS: Over a 17-month period, all children with cancer, fever, and neutropenia admitted to five hospitals in Santiago, Chile, were enrolled onto a prospective protocol. Associations between admission parameters and risk for invasive bacterial infection were assessed by univariate and logistic regression analyses.
RESULTS: A total of 447 febrile neutropenic episodes occurred in 257 children. Five parameters were statistically independent risk factors for an invasive bacterial infection. Ranked by order of significance, they were as follows: C-reactive protein levels of 90 mg/L or higher (relative risk [RR], 4.2; 95% confidence interval [CI], 3.6 to 4.8); presence of hypotension (RR, 2.7; 95% CI, 2.3 to 3.2); relapse of leukemia as cancer type (RR, 1.8, 95% CI, 1.7 to 2.3); platelet count less than or equal to 50,000/mm3 (RR, 1.7; 95% CI, 1.4 to 2.2); and recent ( CONCLUSION: In a large population of children, common clinical and laboratory admission parameters were identified that can help predict the risk for an invasive bacterial infection. These results encourage the possibility of a more selective management strategy for these children.
FEVER AND NEUTROPENIA in cancer patients was considered a medical emergency during the past decade, requiring prompt in-hospital evaluation and administration of broad-spectrum intravenous antibiotics.1 This aggressive approach was universally recommended because approximately 60% of febrile neutropenic episodes were caused by bacterial infections with or without bacteremia, a condition that was difficult to identify in many children at the time of their first evaluation.2,3 Although this aggressive approach reduces mortality,3 patients may suffer adverse effects, such as antimicrobial toxicity, nosocomial infections, fungal superinfections, and the psychologic and financial impact of complex hospital treatment.4 These drawbacks and the recent recognition that these patients form a heterogeneous group at different risks for infection-related morbidity and mortality4 have led clinicians worldwide to consider a more selective management, including a shortened antimicrobial course, early hospital discharge, oral antimicrobial treatment, and management as an outpatient, that would significantly benefit these children and health care systems.5-11 Recent studies have concluded that the critical issue is to find appropriate tools with which to identify those individuals who are at high risk or low risk for an invasive bacterial infection (IBI).12-14 For pediatric patients, studies on identification of risk factors for IBI15-17 and studies on selective management are scarce.18-21 Rackoff et al15 identified two parameters, magnitude of fever and absolute monocyte count (AMC), that proved useful in identifying children at low risk, intermediate risk, or high risk for bacteremia. Other studies in pediatric populations have looked at risk factors associated with overall treatment failure and not with IBI specifically.22 For selective management strategies, Aquino et al18,19 reported that up to two thirds of children with cancer, fever, and neutropenia were at a low risk for bacteremia and could be managed safely with early hospital discharge. These authors estimated that this approach could save $5,000 per patient (United States dollars, here and throughout) and $23,000,000 per year for the United States.20 In a pilot study of outpatient management, this same group successfully treated 19 children at low risk for bacteremia with ceftriaxone.21 Other researchers have reported nonrandomized prospective experiences using intravenous antibiotics in outpatient settings.23,24 In Chile, a rapidly developing country, during the past 10 years, a team of physicians from our Subcommittee of Infectious Diseases of the National Child Program of Antineoplastic Drugs has worked, within an environment of limited resources, in a stepwise manner, to develop more selective strategies for pediatric patients with cancer, fever, and neutropenia. In two consecutive studies, we concluded that serial determination of serum C-reactive protein (CRP) can aid in identifying children at high risk for bacterial infections25 and that for selected children at low-risk for bacteremia, antibiotic treatment can be safely stopped at day 3 of hospitalization.26 In this prospective, multicenter study, our aim was to advance a step further. We studied a significant number of clinical and laboratory variables at enrollment in children with cancer, fever, and neutropenia in order to determine which correlated with the presence of IBI. Our hypothesis was that independent variables significantly associated with a risk for IBI can be identified at enrollment.
Population Selection From May 1,1996, to September 30, 1997, all children 18 years of age or younger receiving cancer chemotherapy who were admitted for fever and severe neutropenia to any of five hospitals were enrolled onto a prospective protocol. The five hospitals were Luis Calvo Mackenna, Exequiel González Cortés, Roberto del Río, San Juan de Dios, and Sótero del Río, all located in Santiago, Chile, and affiliated with the National Health Care System, as well as the University of Chile. Together these hospitals provide health care to two million children within the low-income to middle-income population. The oncology service of each hospital, which is responsible for the management of children with cancer, participates in the National Child Program of Antineoplastic Drugs (PINDA), which establishes standard protocols for treatment of different types of cancer. This study protocol was conducted by physicians from the oncology and infectious disease services of each hospital, under the auspices of the PINDA Subcommittee of Infectious Diseases. Informed consent was obtained for all patients. This study was approved by the ethical committee of each hospital.
Overall Study Design
Evaluation at Enrollment The following entry variables were uniformly recorded: (1) demographic variables, ie, age, sex, and maternal educational level; (2) cancer-related variables, ie, cancer type, intensity of chemotherapy, use of granulocyte colony-stimulating factors since last administration of chemotherapy, and use of an indwelling catheter; (3) variables related to the febrile episode, ie, hours of fever before admission, days since last administration of chemotherapy, and use of prophylactic antimicrobial agents; (4) and admission clinical and laboratory variables, ie, axillary temperature, blood pressure, ANC, AMC, quantitative serum CRP level, hemoglobin level, and platelet count.
Patient Management and Monitoring
Definitions
Statistical Analysis The cutoff points for CRP level and platelet count were determined by constructing a receiver operator characteristic curve.29 All statistical analyses were performed with the BMDP Solo statistical package (SPSS, Inc, Chicago, IL). Results for continuous variables were expressed as mean ± SD.
Overall Description of Febrile Neutropenic Episodes During the 17-month study period, a total of 447 febrile neutropenic episodes occurred in 257 children. Of these children, 148 (58%) had one episode, 63 (25%) had two, 27 (10%) had three, and 19 (7%) had four episodes during the study. Age of children at the time of fever onset ranged from 6 months to 18 years (mean age, 7 ± 4.2 years), sex was male for 50%, and maternal educational level was low for 55%. The characteristics of cancer-related variables at the time of admission for 257 children with 447 febrile neutropenic episodes are listed in Table 1. The predominant cancer type was acute lymphocytic leukemia, and all episodes of fever and neutropenia occurred in children who were receiving chemotherapy. Intensive chemotherapy was being used in 257 episodes (57%). None of the children had received a bone marrow transplant.
IBI Status of Febrile Neutropenic Episodes An IBI based on our definition occurred in 178 (40%) of the 447 episodes. Diagnoses and frequency of microbial confirmation for these 178 episodes are listed in Table 2. According to our definitions, 87 episodes had a demonstrated IBI and 91 had a probable IBI. Bacterial species recovered from 87 episodes of demonstrated IBI are shown in Table 3. Clinical diagnoses in the 91 episodes of probable IBI were based on the following criteria: for bacterial bronchopneumonia, clinical findings and chest radiographs; for sepsis syndrome, see above; for enterocolitis, clinical findings and the presence of blood and/or leukocytes in stools; for typhlitis, clinical findings, abdominal radiographs, and ultrasound; for cellulitis, mucositis, and perianal abscess, clinical findings; and for endocarditis, clinical findings and the presence of vegetations by echocardiography.
Enrollment Variables Associated With Presence of an IBI Presence or absence of an enrollment variable was compared for the episodes associated (all, demonstrated, and probable) and not associated with an IBI (Table 4). Low maternal educational level, presence of a leukemia relapse, use of intensive chemotherapy, presence of an indwelling catheter, an interval of less than or equal to 7 days since last chemotherapy, a temperature of 39°C or higher, presence of hypotension, an ANC of less than or equal to 100/mm3, an AMC of less than or equal to 100/mm3, serum CRP level of 90 mg/L or higher, and platelet count of less than or equal to 50,000/mm3 were all significantly associated with a risk of an IBI by univariate analysis. These variables were included in the logistic regression model.
Five variables proved to be independent risk factors for an IBI (Table 5). These factors, ranked by order of significance, were CRP levels greater than or equal to 90 mg/L, presence of hypotension, relapse of leukemia as a type of cancer, presence of platelet count less than or equal to 50,000/mm3, and an interval of less than or equal to 7 days since last chemotherapy. When analyzed separately for demonstrated and probable IBI groups, CRP, hypotension, and relapse of leukemia remain significant. Variables with lower relative risk (RR), such as low platelet count for the demonstrated IBI group and recent chemotherapy for the probable IBI group, lost significance (Table 5).
Ten children died during the febrile episode, all of them with an IBI (eight were blood culture–positive), for an overall lethality of 10 (3.9%) of 257 patients, or 10 (2.2%) of 447 episodes. All 10 children had two or more risk factors at enrollment, and nine of them had CRP levels of 90 mg/L or higher, eight had hypotension, seven had relapse of leukemia, nine had platelet counts less than or equal to 50,000/mm3, and 10 had an interval of 7 days or less since last chemotherapy.
Predictive Model for the Risk of IBI
Because a second or third febrile neutropenic episode in a same child could eventually influence the model, we also performed the above analysis for the first episode only. For the 257 first febrile episodes, analysis was not statistically different from that for all 447 episodes: CRP levels greater than or equal to 90 mg/L, (RR, 4.0; 95% CI, 3.4 to 4.9), presence of hypotension (RR, 2.5; 95% CI, 2.0 to 3.2), relapse of leukemia as cancer type (RR, 1.8; 95% CI, 1.6 to 2.5), platelet count less than or equal to 50,000/mm3 (RR, 1.5; 95% CI, 1.2 to 2.1), and recent (  7 days) chemotherapy (RR, 1.4; 95% CI, 1.1 to 1.6).
Children with cancer, fever, and neutropenia differ in their risk for IBI, a potentially fatal complication of bone marrow depletion. Previous studies have concluded that children can be separated into different risk categories for IBI according to clinical and laboratory parameters present at the time of admission.15-17 In one study, an AMC less than 100/mm3 and a temperature of 39°C or higher were significantly associated with a higher risk for bacteremia. The 95% CI for the odds ratio comparing high- versus intermediate-risk groups in that study was wide (1.6 to 12.9).15 Two other studies concluded that risk factors were presence of leukemia as the cancer type,16,17 induction chemotherapy treatment, ANC less than or equal to 200/mm3,16 occurrence of chills, hypotension, and need for resuscitation.17 Other studies have been performed in adult populations.30,31 The characteristics of febrile episodes among adults with cancer and neutropenia differ from those that occur in children, mainly in the high proportion of adult patients with comorbidity; therefore, conclusions obtained from these patients cannot be reliably extrapolated to children. In adults with fever, cancer, and neutropenia, serious concurrent comorbidity, such as inadequate fluid intake, inadequate pain control, uncontrolled bleeding, and so on, occurred in 36% of cases.31
In this prospective, multicenter study, we identified five clinical and laboratory parameters that at a time of a first evaluation were independently associated with a higher risk for an IBI. The two variables with strongest association, CRP level 90 mg/L or higher and presence of hypotension, were indicators of significant systemic bacterial infection. In previous studies, we validated the usefulness of serial determinations of serum CRP for the diagnosis and follow-up of bacterial infections (cutoff value of 40 mg/L) in children with cancer, fever, and neutropenia.25,26 This study confirmed that a cutoff value of 90 mg/L is a useful predictor of IBI. The other three variables, relapse of leukemia as cancer type, interval of 7 days or less since last chemotherapy treatment, and platelet count less than or equal to 50,000/mm3, albeit less intensely associated, can be considered patient-related risk factors for an IBI. A more strict analysis by IBI groups defined bacteriologically (demonstrated IBI) and clinically (probable IBI) showed that the variables with higher RR (CRP An analytic strength of our study was the simultaneous assessment of potential risk factors within a large study population allowing for the identification of independent factors predictive of an IBI, as well as for factors that do not add predictive value to an evaluation. In this study, high fever and low AMC, factors that had previously resulted in significant risk factors for IBI,15 reached significance in the univariate but not multivariate analysis. It is possible that a similar situation would have occurred in the study by Rackoff et al if more children had been included. A limitation of our study was the relatively low yield in bacterial detection. A bacterial pathogen was isolated in 20% of all febrile episodes and in 49% of episodes classified as IBI. The relatively low culture yield does not differ from that reported in other publications and reflects our current diagnostic capacity.12,13,15 Our principal aim was to identify independent factors that predict risk of IBI in pediatric cancer patients with fever and neutropenia, a requirement for establishing a model that can be safely adopted to manage these children. The variables have been identified; children that present with absence of risk factors are at a low risk (2%) of having an IBI. These children are evident candidates for outpatient management. On the other hand, children with two or more risk factors have a risk of IBI that surpasses 48%; these children will require a more aggressive approach. The group that is more complex includes 33% of children who have only one risk factor. With these results, we can propose a model of risk prediction that takes into account the number and type of risk factors present at enrollment. A child with a febrile neutropenic episode can be considered at high risk for IBI if he or she has two or more risk factors or a serum CRP level of 90 mg/L or higher, hypotension, or relapse of leukemia as sole factors; the child can be considered at low risk if he or she has an absence of risk factors, a platelet count less than or equal to 50,000/mm3, or recent chemotherapy as sole factors. In this study, this model would have a sensitivity of 90%, specificity of 65%, and positive and negative predictive values of 75% and 87%, respectively. A prospective validation of this model is required before we can propose that it be used for the adoption of a selective management strategy. We are currently validating the predictive value of this model in a prospective, government-sponsored, multicenter study. A strategy that considers a selective approach for febrile neutropenic children will be of significant benefit to the increasing population of children with cancer and also to the health care systems of industrialized and developing countries worldwide.
1. Pizzo PA, Rubin M, Freifeld A, et al: The child with cancer and infection: I. Empiric therapy for fever and neutropenia, and preventive strategies. J Pediatr 119: 679-694, 1991[Medline] 2. Klastersky J: Febrile neutropenia. Curr Opin Oncol 5: 625-632, 1993[Medline] 3. Pizzo PA, Robichaud KJ, Wesley R: Fever in the pediatric and young adult patient with cancer: A prospective study of 1001 episodes. Medicine 61: 163-165, 1982 4. Anaissie E, Vadhan-Raj S: Is it time to redefine the management of febrile neutropenia in cancer patients? Am J Med 98: 221-223, 1995[Medline] 5. Buchanan G: Approach to treatment of the febrile cancer patient with low-risk neutropenia. Hematol Oncol Clin North Am 7: 919-935, 1993[Medline] 6. Cohen KJ, Leamer K, Odom L, et al: Cessation of antibiotics regardless of ANC is safe in children with febrile neutropenia: A preliminary prospective trial. J Pediatr Hematol Oncol 17: 325-330, 1995[Medline] 7. Malik IA, Abbas Z, Karim M: Randomized comparison of oral ofloxacin alone and in combination of parenteral antibiotics in neutropenic febrile patients. Lancet 339: 1092-1096, 1992[Medline] 8. Rubenstein EB, Rolston K: Outpatient treatment of febrile neutropenic patients with cancer. Eur J Cancer 31: 2-4, 1995 9. Freifeld AG, Pizzo PA: The outpatient management of febrile neutropenia in cancer patients. Oncology 10: 599-606, 1996[Medline] 10. 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] 11. Escalante CP, Rubenstein EB, Rolston KV: Outpatient antibiotic therapy for febrile episodes in low-risk neutropenic patients with cancer. Cancer Invest 15: 237-242, 1997[Medline]
12.
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
13.
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
14.
Finberg R, Talcott J: Fever and neutropenia: How to use a new treatment strategy. N Engl J Med 341: 362-363, 1999 (editorial)
15.
Rackoff WR, Gonin R, Robinson C, et al: Predicting the risk of bacteremia in children with fever and neutropenia. J Clin Oncol 14: 919-924, 1996 16. Jones GR, Konsler GK, Dunaway RP, et al: Infection risk factors in febrile, neutropenic children and adolescents. Pediatr Hematol Oncol 13: 217-229, 1996[Medline] 17. 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] 18. Aquino VM, Tkaezewski Y, Buchanan G: Early discharge of low-risk febrile neutropenic children and adolescents with cancer. Clin Infect Dis 25: 74-78, 1997[Medline] 19. Aquino VM, Buchanan GR, Tkaczewsky Y, et al: Safety of early hospital discharge of selected febrile children and adolescents with cancer with prolonged neutropenia. Med Pediatr Oncol 28: 191-195, 1997[Medline] 20. Bash R, Katz J, Cash J, et al: Safety and cost effectiveness of early hospital discharge of lower risk children with cancer admitted for fever and neutropenia. Cancer 74: 189-196, 1994[Medline] 21. 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] 22. Miranda-Novales MG, Belmont-Martinez L, Villasis-Keever M: Empirical antimicrobial therapy in pediatrics patients with neutropenia and fever: Risk factors for treatment failure. Arch Med Res 4: 331-335, 1998 23. Kaplinsky C, Druker M, Goshen J, et al: Ambulatory treatment with ceftriaxone in febrile neutropenic children. Isr J Med Sc 30: 649-651, 1994[Medline] 24. Shemesh E, Yaniv I, Drucker M: Home intravenous antibiotic treatment for febrile episodes in immunocompromised pediatric patients. Med Pediatr Oncol 30: 95-100, 1998[Medline] 25. Santolaya ME, Cofré J, Beresi V: C-reactive protein: Valuable aid for the management of febrile children with cancer and neutropenia. Clin Infect Dis 18: 589-595, 1994[Medline] 26. Santolaya ME, Villarroel M, Avendaño L, et al: Discontinuation of antimicrobial therapy for febrile, neutropenic children with cancer: Prospective study. Clin Infect Dis 25: 92-97, 1997[Medline] 27. Sáez Llorens X, McCracken GH: Sepsis syndrome and septic shock in pediatrics: Current concepts of terminology, pathophysiology, and management. J Pediatr 123: 497-508, 1993[Medline] 28. Vargas L, Becker A, Beresi V, et al: Resultados por protocolos PINDA, in Cancer Infantil en Chile: Diez a2 ños del Programa PINDA. Santiago, Chile, Ministerio de Salud de Chile, 2000, pp 161-268 29. Sackett DL, Haynes RB, Tugwell P: Clinical Epidemiology: A Basic Science for Clinical Medicine, Chapter 9 : Diagnostic test. Boston, MA, Little Brown, 1985, pp 101-106 30. Talcott J, Finberg R, Mayer R, et al: The medical course of cancer patients with fever and neutropenia. Arch Intern Med 148: 2561-2568, 1988[Abstract] 31. Talcott J, Siegel R, 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[Abstract] Submitted January 22, 2001; accepted April 17, 2001. This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2001 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
90 mg/L, hypotension, and relapse of leukemia) remained significant despite separate analysis. Variables with lower RR maintained the trend (although low platelet count for demonstrated IBI and recent chemotherapy for probable IBI lost significance). This may have been more a problem of sample size after splitting of the analysis than a true loss of significance, an issue that will require further evaluation. 
