Oncologist. The. Symptom Management and Supportive Care. Risk Models for Predicting Chemotherapy-Induced Neutropenia

Transcription

1 The Oncologist Symptom Management and Supportive Care Risk Models for Predicting Chemotherapy-Induced Neutropenia GARY H. LYMAN, CHRISTOPHER H. LYMAN, OLAYEMI AGBOOLA, FOR THE ANC STUDY GROUP Health Services and Outcomes Research Program, James P. Wilmot Cancer Center, University of Rochester School of Medicine and Dentistry, Rochester, New York, USA ABSTRACT Key Words. Neutropenia Toxicities Colony-stimulating factors Risk model Chemotherapy Neutropenia and its complications, including febrile neutropenia, are major dose-limiting toxicities of systemic cancer chemotherapy. A number of studies have attempted to identify risk factors for neutropenia and its consequences to develop predictive models capable of identifying patients at greater risk for such complications and to guide more effective and costeffective applications of the colony-stimulating factors. A systematic review of the literature showed that age, performance status, nutritional status, chemotherapy dose intensity, and low baseline blood cell counts were associated with the risk of severe and febrile neutropenia or reduced chemotherapy dose intensity in multivariate analysis in two or more studies. Similarly, age, diagnosis of leukemia or INTRODUCTION Chemotherapy-induced neutropenia (CIN) is the major dose-limiting toxicity of systemic cancer chemotherapy, and it is associated with substantial morbidity, mortality, and costs. Neutropenia may result in fever and neutropenia, or febrile neutropenia (FN), often necessitating hospitalization for evaluation and empiric broad-spectrum antibiotics. Such complications often result in dose reductions or treatment delays, which may compromise clinical outcomes [1 5]. The prophylactic use of colony-stimulating factors (CSFs) can reduce the risk, severity, and duration of both severe and febrile neutropenia [6 11]. Despite lymphoma, high temperature or low blood pressure at admission, and i.v. site infection along with low blood cell counts and organ dysfunction were associated with serious medical complications of febrile neutropenia, including bacteremia and death. The available risk model studies, however, had several limitations, including retrospective analyses of small study populations lacking independent validation, frequent missing values, and differences in the predictive factors considered. To overcome the limitations of previous studies, efforts are under way to develop and validate risk models based on large prospective studies in representative populations of patients receiving systemic chemotherapy. The Oncologist 2005;10: these benefits, CSFs are not administered to all patients receiving myelosuppressive chemotherapy because of the costs associated with their routine use. The selective use of CSFs in patients at greater risk for neutropenic complications may, however, enhance their cost-effective use by directing treatment toward those patients who are most likely to benefit. The risk of severe and febrile neutropenia is usually based on the treatment regimen. A survey of the literature on randomized clinical trials of chemotherapy in patients with early-stage breast cancer (ESBC) and non-hodgkin s lymphoma (NHL) has shown, however, that the rates of Correspondence: Gary H. Lyman, M.D., M.P.H., F.R.C.P. (Edin), University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue, Box 704, Rochester, New York 14642, USA. Telephone: ; Fax: ; Gary_Lyman@urmc.rochester.edu Received January 3, 2005; accepted for publication March 31, AlphaMed Press /2005/$12.00/0 The Oncologist 2005;10:

2 428 Risk Models for Chemotherapy-Induced Neutropenia myelosuppression and relative dose intensity (RDI) are underreported. When reported, the rates with the same and similar regimens varied greatly, making it difficult to determine the actual risk for neutropenic complications associated with common chemotherapy regimens [12]. Differences in the reported rates of neutropenic complications may relate to differences in study patient populations as well as the delivered dose intensity. Treatment dose intensity was reported with even less consistency, making it very difficult to interpret differences in reported rates of toxicity or treatment efficacy. Early estimates of the threshold of risk at which the prophylactic use of CSFs becomes cost saving have been provided by economic models based on cost minimization using itemized, fixed costs of hospitalization for FN. These early models showed that the use of CSFs was cost saving when the risk for hospitalization for FN exceeded 40%; later models that are based on more current, direct costs and that include some indirect costs have shown that appropriate thresholds are closer to 20% [13, 14]. Clearly, however, models based on average population risk estimates and the presumed myelosuppressive potential of the chemotherapy regimen may not be the most effective way for determining which patients should be offered CSF support. Therefore, recent efforts to increase the cost-effectiveness of the use of CSFs have focused on identifying patients at greater risk for neutropenic events, who are most likely to benefit from CSF prophylaxis. The development of clinical prediction or risk models for neutropenia and its complications based on risk Table 1. Risk factors for neutropenia or leukopenia by study: multivariate model results (odds ratios) factors should enable the identification of patients at greatest risk for these complications and the more targeted and costeffective application of appropriate supportive care measures, such as the CSFs, in those at greatest need, who are most likely to benefit [15]. To identify risk factors associated with neutropenia and its complications, a systematic review of the literature was undertaken using electronic databases, including MED- LINE, EMBASE, and the Cochrane Library as well as references from relevant articles that were identified. Search terms included neutropenia, febrile neutropenia, risk factors, risk models, and predictive models. The studies that were selected reported multivariate risk models of neutropenic events or complications as well as RDI in patients receiving systemic cancer chemotherapy. In these studies, neutropenia was generally defined as an absolute neutrophil count (ANC) < /l (grade 3 and 4) or ANC < /l (grade 4). FN was defined as severe neutropenia (grade 3 or higher) with temperature >38.2 C. Dose intensity was defined as dose administered per unit of time, for example, mg/m 2 per week. In turn, the average RDI was defined as the ratio of the administered dose intensity to the standard dose intensity averaged across drugs in a regimen. RISK FACTORS FOR CIN Studies reviewed in this article showed risk factors for CIN or leukopenia (Table 1), FN (Table 2), and RDI (Table 3) in patients with NHL [16 22], ESBC [22 26], and other malignancies [27 29]. In these studies, advanced age, female sex, Voog et al. Intragumtornchai et al. Kloess et al. Aslani et al. Blay et al. (2000) [16] (2000) [17] (conditional) (1999) [18] (2000) [23] (1996) [29] Malignancy NHL NHL NHL Breast cancer Various cancers Validation No No Yes No Yes Cyclophosphamide (HD) 19.8 High TNF receptor level 8.52 Poor performance status Older age Nitrogen index level < No CSF use High-risk chemotherapy 4.76 Day 5 lymphocyte count < /l 6.17 LDH level > ULN Bone marrow involvement 5.8 Albumin level <3.5 g/dl 16.2 Female sex 2.0 Etoposide use 6.1 Abbreviations: CSF = colony-stimulating factor; HD = high dose; LDH = lactate dehydrogenase; NHL = non-hodgkin s lymphoma; TNF = tumor necrosis factor; ULN = upper limit of normal.

3 Lyman, Lyman, Agboola et al. 429 Table 2. Risk factors for febrile neutropenia by study: multivariate model results (odds ratios) Voog et al. Intragumtornchai Lyman et al. Morrison Lyman et al. Morrison Ray-Coquard Blay et al. (2000) et al. (2000) [17] (2003) et al. (2001) (2004) et al. (2004) et al. (2003) (1996) [16] (conditional) [19] [20] [21] [22] [28] [29] Malignancy NHL NHL NHL NHL NHL Breast Various Various cancer cancers cancers and NHL Validation No No No No No No No Yes Cyclophosphamide (HD) 14.7 Poor performance status High TNF receptor level 3.49 High-risk chemotherapy Lymphocyte count < /l LDH level > ULN 3.1 Bone marrow involvement 4.9 Albumin level <3.5 g/dl 11.2 Older age Female sex 1.32 Renal disease Heart disease 1.54 Heart, renal, or liver disease 2.12 Hemoglobin level 12 g/dl Planned full chemotherapy 2.41 a 2.7 b 2.74 dose intensity No CSF use NHL 1.64 Advanced-stage disease 1.30 Pretreatment ANC < /l 1.53 a RDI >80%. b RDI 80%. Abbreviations: ANC = absolute neutrophil count; CSF = colony-stimulating factor; HD = high dose; LDH = lactate dehydrogenase; NHL = non-hodgkin s lymphoma; RDI = relative dose intensity; TNF = tumor necrosis factor; ULN = upper limit of normal. poor performance status, poor nutritional status, and low baseline and first-cycle nadir blood cell counts along with high chemotherapy dose intensity were all significant predictors of neutropenic events (Fig. 1). In patients with NHL, a high lactate dehydrogenase (LDH) level and the presence of bone marrow involvement were also significant. Studies that assessed risk factors for the consequences of neutropenia, including serious medical complications and mortality, were in patients with solid tumors, leukemias, and lymphomas in adults [30 38] and in children [39 43]. Significant predictors for neutropenic complications including death (Table 4), bacteremia (Table 5), and length of hospital stay 10 days (Table 6) were advanced age, hematologic malignancies, greater disease burden, high temperature and low blood pressure on admission, pneumonia, i.v. site infection, low blood cell counts, and organ dysfunction. The common risk factors for CIN and its complications can be classified based on disease, patient characteristics, and type of treatment. Disease-Specific Risk Factors Tumor Type Patients with hematologic malignancies are at greater risk for neutropenic complications than patients with solid tumors because of the underlying disease process as well as the intensity of the treatment that is required [30, 31]. Consistent with these findings, an analysis of data from patients hospitalized with FN in the 120 hospitals of the University HealthSystem Consortium showed that the duration of hospitalization as well as inpatient mortality were significantly greater in patients with leukemia than in patients with solid tumors (p <.001 for both) [32].

4 430 Risk Models for Chemotherapy-Induced Neutropenia Table 3. Risk factors for reduced dose intensity by study: multivariate model results (odds ratios) Silber et al. (1998) Lyman et al. Savvides et al. Szucs et al. Morrison et al. Lyman et al. [27] (conditional) (2003) [24] (2003) [25] (2004) [26] (2001) [20] (2004) [21] Malignancy Breast cancer Breast cancer Breast cancer Breast cancer NHL NHL Validation Yes No Yes No No No Reduced relative dose intensity ( 80% 85%) Older age Body surface area >2 m CMF chemotherapy 3.4 CAF chemotherapy day cycle 2.5 Heart disease 1.8 First-cycle ANC nadir 0.14 Percent drop platelet count 1.04 Neutropenic events 4.9 Nonanthracycline regimen 1.5 Concomitant radiotherapy 1.3 No CSF use in patients >60 years 2.41 Advanced-stage disease 1.18 Poor performance status 1.28 CSF use 0.70 Concurrent radiotherapy 9.48 Cycle 1 ANC < /l 4.4 Hemoglobin level drop 1.8 Composite neutropenia and/or reduced dose intensity Abbreviations: ANC = absolute neutrophil count; CAF = cyclophosphamide, doxorubicin, and fluorouracil; CMF = cyclophosphamide, methotrexate, and fluorouracil; CSF = colony-stimulating factor; NHL = non-hodgkin s lymphoma. Age Female sex Performance status Nutrition Body surface area Temperature Blood pressure Infection at intravenous site Pneumonia Antifungal prophylaxis No antibiotics The course of neutropenia and its complications Chemotherapy intensity Neutropenia Febrile neutropenia Complicated infection Bacteremia Prolonged hospitalization Death Pretreatment blood cell counts Lymphocyte counts Hemoglobin level Lactate dehydrogenase level Bone marrow involvement Neutrophil counts Prolonged neutropenia Monocyte counts Platelet counts Blood urea nitrogen level Glucose level Lactate dehydrogenase level Burden of illness Patient-Specific Risk Factors Age Ten studies found higher age to be a general risk factor for the development of severe neutropenia [18 24] and other neutropenic complications [30, 32, 33]. Advanced age is a particularly important independent risk factor, since older patients are often treated with lower chemotherapy doses to minimize the occurrence of neutropenic complications. Since older patients with cancer can obtain the same benefit from aggressive chemotherapy as younger patients [44], effective management of the risk of neutropenia is crucial to make it possible to administer full-dose chemotherapy in this population. Figure 1. Risk factors for chemotherapy-induced neutropenia and its complications identified from published clinical prediction models. Advanced Disease and Uncontrolled Cancer Both advanced disease and uncontrolled cancer were significant predictors of RDI, hospitalization for FN, and serious neutropenic complications, including death [21, 26, 33, 34]. Performance Status Three studies have shown that, in addition to age, poor performance status is a significant risk factor for chemotherapy-induced neutropenia [16, 18, 27]. In older patients, physiologic age or frailty may be a more accurate predictor of risk than chronologic age [45].

5 Lyman, Lyman, Agboola et al. 431 Table 4. Risk factors for death by study: multivariate model results (odds ratios) Kim et al. Leong et al. Basu et al. Kuderer et al. Kuderer et al. Gonzalez-Barca Study (2000) [33] a (1997) [31] b (2004) [43] c (2002) [32] (2004) [38] et al. (1999) [34] Malignancy Solid tumors Solid tumors Various cancers Various cancers Various cancers Various cancers or lymphomas with bacteremia Validation No No No No No No Severe burden of illness 5.77 Complex infection 4.31 Lymphoma 3.73 Low platelet count 1.64 Shock 10.0 Pneumonia Uncontrolled cancer ANC < /l and initial regimen failure No norfloxacin use 2.4 Length of hospital stay >5 days 6.86 Infant age 2.06 Adolescent age 2.26 Hematologic malignancy 1.48 Leukemia Comorbidities Advanced age Male sex Secondary febrile neutropenia Gram-negative sepsis 4.48 Gram-positive sepsis 2.28 Lung disease 3.95 Renal disease 3.19 Liver disease 2.93 Cerebrovascular disease 3.29 a Outcomes also included other complications of febrile neutropenia. b Outcomes also included prolonged neutropenia or hospitalization. c Pediatric study. Abbreviation: ANC = absolute neutrophil count. Comorbidities The presence of comorbid conditions with cancer has been shown to increase the risk for neutropenia and its complications. Renal disease and heart disease have been shown to increase the risk for FN and reduced RDI in patients with NHL [19, 20]. In patients with breast cancer, liver disease along with kidney and heart disease increased the risk for severe neutropenia and FN [22]. Similarly, comorbidities such as hypertension, chronic obstructive pulmonary disease, pneumonia, prior fungal infection, and sepsis have been shown to increase the risk for serious neutropenic complications, including prolonged hospitalizations for FN and death [30, 32, 34, 38]. Laboratory Abnormalities Blood cell counts and chemistry analyses may indicate either the extent of disease, comorbid conditions, or the impact of cytotoxic chemotherapy. Many of these laboratory abnormalities have been identified as predictors for CIN and its complications. We have reported that pretreatment WBCs are predictive of both FN and an RDI <85% in patients with ESBC [22]. Pretreatment hemoglobin levels <12 g/dl are also predictors of severe neutropenia or FN in cycle 1 [19, 22]. One study in patients with aggressive NHL treated with cyclophosphamide, doxorubicin (Adriamycin ; Bedford Laboratories, Bedford, OH, vincristine

6 432 Risk Models for Chemotherapy-Induced Neutropenia Table 5. Risk factors for bacteremia by study: multivariate model results (odds ratios) Viscoli et al. Rackoff et al. Hann et al. Santolaya et al. Klaassen et al. Study (1994) [36] (1996) [39] a (1997) [40] a,b (2001) [41] a,b (2000) [42] a Malignancy Leukemia and Various cancers Various cancers Various cancers Various cancers solid tumors Validation Yes No No Yes Yes Antifungal prophylaxis 2.48 Granulocytopenia >16 days duration 5.37 Low platelet count 0.51 c 1.7 d Temperature at admission > 40 C 8.41 Shock Infection at i.v. site 3.32 Quinolone prophylaxis 0.47 Absolute monocyte count < /l No hematologic malignancy 0.47 Presence of i.v. line 1.87 Neutropenia before fever >15 days duration 7.64 Higher temperature 2.35 CRP level 90 mg/l 4.2 Hypotension 2.7 Relapsed leukemia 1.8 <7 days since chemotherapy 1.3 AML or NHL 1.9 Bone marrow involvement 3.7 Unwell appearance 2.3 a Pediatric study. b Relative risk. c < /l. d /l. Abbreviations: AML = acute myelogenous leukemia; CRP = C-reactive protein; NHL = non-hodgkin s lymphoma. Table 6. Risk factors for length of stay 10 days by study: multivariate model results (odds ratios) Kuderer et al. Kuderer et al. (2002) [32] (2004) [38] Malignancy Various cancers Various cancers Validation No No Leukemia Complications Age 65 years Secondary febrile neutropenia Gram-positive sepsis 2.39 Renal disease 2.22 Lung disease 2.76 Cerebrovascular disease 2.26 Venous thrombosis 2.21 Liver disease 2.14 Multiple cancers 2.14 Infection at i.v. site 2.02 (Oncovin ; Eli Lilly and Company, Indianapolis, IN, and prednisone (Deltasone ; Pfizer Pharmaceuticals, New York, (CHOP) also found that a serum albumin concentration of 35 g/l or less, an LDH level greater than the upper limit of normal, and bone marrow involvement of the lymphoma were all significant predictors of life-threatening neutropenia and FN [17]. The predictive value of high LDH levels was also determined in another study in patients with NHL, by Kloess et al. [18]. Conditional risk models have used patient responses in the first cycle of chemotherapy to predict the occurrence of neutropenia and its complications in subsequent cycles. Three studies have shown the predictive value of the first-cycle nadir in leukocyte counts [25, 27, 29] and decreases in hemoglobin or platelet levels [25, 27] for predicting neutropenic complications in later cycles. An ANC < /l is also a significant predictor of serious medical complications [33].

7 Lyman, Lyman, Agboola et al. 433 Treatment-Specific Risk Factors Chemotherapy Regimen The intensity of specific chemotherapy regimens is one of the primary determinants of the risk for CIN, with some regimens being more myelosuppressive than others [29]. High cyclophosphamide doses or the use of etoposide (Etopophos, VePesid ; Bristol-Myers Squibb, Princeton, NJ, bms.com) in treating patients with NHL [16, 18] and high anthracycline doses in treating patients with ESBC [24, 26] have all been identified as significant predictors for severe neutropenia and FN. High chemotherapy dose intensity has also been identified as a risk factor for neutropenia and its complications in older patients with NHL [19, 20, 22, 26]. CSF Use Five studies in lymphoma have shown that prophylactically administered CSF is associated with a significantly lower risk of severe neutropenia and FN [16, 18, 20 21]. Predictors of Neutropenic Complications in Children Most of the studies summarized above were in adult patients with cancer, but predictors of neutropenic complications have also been found in pediatric patients with acute lymphoblastic leukemia and solid tumors (Table 4 and Table 5) [39, 40]. In pediatric patients, higher temperature, prolonged neutropenia, a low monocyte count, and shock predicted the occurrence of bacteremia [39, 40, 42]. RISK FACTORS FOR CIN AND GUIDELINES FOR CSF USE Several of the risk factors identified in this study have been incorporated into current guidelines for CSF use. For example, the 2000 update of the American Society of Clinical Oncology (ASCO) guidelines recommended that CSF be initiated in the first cycle in patients at high risk for FN due to special circumstances, such as patients with poor performance status scores, marrow compromise, advanced cancer, overall immune deficit, or active ongoing infections [1]. The greater risk for CIN in older patients was also acknowledged by the Senior Adult Oncology Panel of the National Comprehensive Cancer Network (NCCN) and the Elderly Task Force of the European Organization for Research and Treatment of Cancer (EORTC). Both organizations recommend that prophylactic CSF be used in patients 65 years of age (EORTC defines elderly patients as 60 years) and in patients with lymphoma who are treated with CHOP or CHOP-like chemotherapy [46, 47]. RISK MODELS VERSUS RISK FACTORS Studies of individual risk factors for neutropenia and its complications, including reduced RDI, provide useful information for identifying patients at greater risk. However, multivariate risk models capable of considering several independent risk factors are a more powerful method for identifying patients at greater risk for these events. In an analysis of data on 577 patients with intermediate-grade NHL, six independently significant risk factors for FN were identified: age 65 years, renal disease, cardiovascular disease, pretreatment hemoglobin level <12 g/dl, a planned average RDI 80%, and no CSF prophylaxis [19]. Risk scores were calculated for each patient based on the number of risk factors. The risk of FN was found to be lower in patients who had fewer than three risk factors than in patients who had three or more, with the greatest risk occurring in the first cycle of chemotherapy (Fig. 2) [19]. CONDITIONAL VERSUS UNCONDITIONAL MODELS Patients in whom FN has occurred are at greater risk for neutropenic complications in subsequent cycles. This has led to interest in conditional models that include the results from the initial cycle as independent variables or predictors [27, 29, 33]. Neutropenic complications are most likely to occur in the first cycle of chemotherapy [19, 48]. This is well established for patients with NHL treated with CHOP chemotherapy (Fig. 3). For this reason, current research has Cumulative probability of FN n of risk factors 5 or or 1 Risk Time (days) Days after CHOP Figure 2. Cumulative probability of febrile neutropenia, by number of risk factors. Figure 3. Risk of first episode of febrile neutropenia in patients with non-hodgkin s lymphoma treated with CHOP chemotherapy.

8 434 Risk Models for Chemotherapy-Induced Neutropenia focused on developing a reliable unconditional model, dependent on pretreatment information alone. Hopefully, such models will be able to guide the early targeted use of supportive care measures in appropriate patients before most complications occur. VALIDATION OF RISK MODELS Four of the 14 models for CIN reviewed in this study have been validated in a group of patients independent from those used for developing the models [18, 25, 27, 29]. Silber et al. developed both conditional and unconditional models to predict severe neutropenia [27]. They validated both models in a separate group of 80 patients with ESBC treated with adjuvant chemotherapy and found that the unconditional model was not reliable in predicting neutropenia. In the conditional model, however, the depth of the ANC nadir in the first chemotherapy cycle predicted neutropenic events in subsequent chemotherapy cycles. LIMITATIONS OF CURRENT RISK MODELS The clinical prediction models that have been developed to date have been based on retrospective data with many important methodologic limitations [15]. As summarized in Table 7, a good clinical model should minimize bias by clearly stating the hypothesis or purpose in advance, including Table 7. Questions for interpreting the findings of risk model studies any subgroup analysis; clearly defining the study population, with specific inclusion and exclusion criteria; and limiting missing data. Models should also be validated in a different set of patients from those in which they were developed. The studies reviewed here are all retrospective in nature, and most have not been designed to test a welldefined hypothesis. They have also been conducted in relatively small study populations and have considered and adjusted for different variables. Furthermore, most models have not been independently validated, and of those models that have been tested in an independent population, only the conditional models are reliable in their predictive power [18, 25, 27, 29]. RISK MODELS DEVELOPED FROM A PROSPECTIVE REGISTRY To overcome these limitations, we have recently implemented a prospective registry of patients initiating a systemic cancer chemotherapy regimen for one of several tumor types. This study involves the comprehensive collection of patient data, including cycle-by-cycle information on hematologic function, neutropenic events, and chemotherapy dosing record, together with patient comorbidities and performance status. The data collected by the registry will be used to develop a multivariate risk model that will Were hypotheses explicitly specified prior to the study? Were missing data limited, that is, less than 10%? Were the analyses based on single occurrences in separate patients rather than on repeated occurrences in the same patient? Were the issues of multiple comparisons considered when many prognostic factors or cut points were evaluated and were the tests of significance adjusted accordingly? Were the problems associated with stepwise multiple regression models considered, including model instability, exaggeration of coefficient estimates, and their significance? Was the effect of any new prognostic factor adjusted for the effects of known prognostic factors? Were the differences in outcomes among subgroups assessed by testing for interaction between the prognostic factor and the variable that defined the subgroups rather than by separate analyses within the subgroups? Were the issues of multiple testing or small sample sizes within subgroups considered when apparent differences among subgroups were interpreted? Was the analysis of subgroups defined only during or after the completion of the study acknowledged as exploratory? Was the prognostic model generated in the developmental dataset validated in an independent dataset of similar patients? In reporting the results of a prognostic factor study: Were the following clearly presented: study design (exploratory or confirmatory, prospective or retrospective), treatment (e.g., randomized or standardized), blinding, main outcomes? Was the number of patients who were excluded because of missing data stated? Was the study duration, including the criteria for study termination, stated? Were the methods of measuring prognostic markers, along with information about reproducibility, stated? Were all study end points clearly defined? Were the data on outcomes summarized as quantitative estimates with confidence intervals? Were the differences in outcomes observed in all patients emphasized more than those found in subgroups? Were any weaknesses of the study, especially those related to subgroup analyses and multiple comparisons, discussed?

9 Lyman, Lyman, Agboola et al. 435 include many of the clinically relevant variables discussed in this review [49 52]. The risk models will be built on the basis of pretreatment patient and disease characteristics as well as planned and delivered chemotherapy. The predictive accuracy of these models will be validated in an independent patient population. Additional studies will evaluate the effect of model-driven treatment interventions on shortterm outcomes, such as neutropenic complications, infection-related mortality, and chemotherapy dose intensity, as well as on long-term outcomes, such as disease-free and overall survival and patient quality of life. This registry should thus make it possible to develop more accurate risk models that can eventually be incorporated into actual clinical practice. Such models should facilitate the targeted application of supportive care measures toward patients at greatest risk and most likely to benefit, therefore reducing the risk of neutropenic events, improving treatment outcomes, and reducing health care costs. CONCLUSIONS Many risk factors for CIN and its serious medical complications have been identified, and several risk models have been developed. The multivariate risk models discussed here have identified advanced age, poor performance status, REFERENCES 1 Ozer H, Armitage JO, Bennett CL et al update of recommendations for the use of hematopoietic colony-stimulating factors: evidence-based, clinical practice guidelines. American Society of Clinical Oncology Growth Factors Expert Panel. J Clin Oncol 2000;18: Rolston KV. New trends in patient management: risk-based therapy for febrile patients with neutropenia. Clin Infect Dis 1999;29: Bonadonna G, Valagussa P, Moliterni A et al. Adjuvant cyclophosphamide, methotrexate, and fluorouracil in nodepositive breast cancer: the results of 20 years of follow-up. N Engl J Med 1995;332: Budman DR, Berry DA, Cirrincione CT et al. Dose and dose intensity as determinants of outcome in the adjuvant treatment of breast cancer. The Cancer and Leukemia Group B. J Natl Cancer Inst 1998;90: Kwak LW, Halpern J, Olshen RA et al. Prognostic significance of actual dose intensity in diffuse large-cell lymphoma: results of a tree-structured survival analysis. J Clin Oncol 1990;8: Crawford J, Ozer H, Stoller R et al. Reduction by granulocyte colony-stimulating factor of fever and neutropenia induced by chemotherapy in patients with small-cell lung cancer. N Engl J Med 1991;325: Trillet-Lenoir V, Green J, Manegold C et al. Recombinant granulocyte colony stimulating factor reduces the infectious comorbidities, and low baseline blood cell counts, along with high chemotherapy dose intensity, as significant predictors for neutropenic complications including FN and RDI. Independent risk factors for adverse consequences of neutropenia, including serious medical complications or death, include advanced age, hematologic malignancies, high temperature and low blood pressure on admission, pneumonia, and i.v. site infection, as well as low blood cell counts and evidence of organ dysfunction. The use of such risk factors to identify the patients who are at greatest risk for neutropenia, and targeting prophylactic CSF to those patients, should help providers focus resources on those who are most at risk. The utility of these models is limited, however, by their retrospective design, the small populations studied, and the various risk factors and methods used. To overcome these limitations, we have designed a prospective registry in different tumor types, which should make it possible to develop accurate and valid risk models for CIN and its complications. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST G. Lyman is on the Speaker s Bureau for Ortho Biotech and Amgen and receives research funding from Amgen, GlaxoSmithKlein, and Genomic Health. The ANC Study Group receives research support from Amgen. complications of cytotoxic chemotherapy. Eur J Cancer 1993;29A: Pettengell R, Gurney H, Radford JA et al. Granulocyte colony-stimulating factor to prevent dose-limiting neutropenia in non-hodgkin s lymphoma: a randomized controlled trial. Blood 1992;80: Holmes FA, O Shaughnessy JA, Vukelja S et al. Blinded, randomized, multicenter study to evaluate single administration pegfilgrastim once per cycle versus daily filgrastim as an adjunct to chemotherapy in patients with high-risk stage II or stage III/IV breast cancer. J Clin Oncol 2002;20: Green MD, Koelbl H, Baselga J et al. A randomized doubleblind multicenter phase III study of fixed-dose single-administration pegfilgrastim versus daily filgrastim in patients receiving myelosuppressive chemotherapy. Ann Oncol 2003;14: Dale DC. Colony-stimulating factors for the management of neutropenia in cancer patients. Drugs 2002;62(suppl 1): Dale DC, McCarter GC, Crawford J et al. Chemotherapyinduced neutropenia and associated complications in randomized clinical trials: an evidence-based review. JNCCN 2003;1: Lyman GH, Kuderer N, Greene J et al. The economics of febrile neutropenia: implications for the use of colony-stimulating factors. Eur J Cancer 1998;34:

10 436 Risk Models for Chemotherapy-Induced Neutropenia 14 Cosler LE, Calhoun EA, Agboola O et al. Effects of indirect and additional direct costs on the risk threshold for prophylaxis with colony-stimulating factors in patients at risk for severe neutropenia from cancer chemotherapy. Pharmacotherapy 2004;24: Lyman GH. A predictive model for neutropenia associated with cancer chemotherapy. Pharmacotherapy 2000;20:104S 111S. 16 Voog E, Bienvenu J, Warzocha K et al. Factors that predict chemotherapy-induced myelosuppression in lymphoma patients: role of the tumor necrosis factor ligand-receptor system. J Clin Oncol 2000;18: Intragumtornchai T, Sutheesophon J, Sutcharitchan P et al. A predictive model for life-threatening neutropenia and febrile neutropenia after the first course of CHOP chemotherapy in patients with aggressive non-hodgkin s lymphoma. Leuk Lymphoma 2000;37: Kloess M, Wunderlich A, Trümper L et al. Predicting hematotoxicity in multicycle chemotherapy. Blood 1999;94(suppl 1):87a. 19 Lyman GH, Morrison VA, Dale DC et al. Risk of febrile neutropenia among patients with intermediate-grade non-hodgkin s lymphoma receiving CHOP chemotherapy. Leuk Lymphoma 2003;44: Morrison VA, Picozzi V, Scott S et al. The impact of age on delivered dose intensity and hospitalizations for febrile neutropenia in patients with intermediate-grade non-hodgkin s lymphoma receiving initial CHOP chemotherapy: a risk factor analysis. Clin Lymphoma 2001;2: Lyman GH, Dale DC, Friedberg J et al. Incidence and predictors of low chemotherapy dose-intensity in aggressive non-hodgkin s lymphoma: a nationwide study. J Clin Oncol 2004;22: Morrison VA, Caggiano V, Fridman M et al. A model to predict chemotherapy-related severe or febrile neutropenia in cycle one among breast cancer and lymphoma patients. Proc Am Soc Clin Oncol 2004;23: Aslani A, Smith RC, Allen BJ et al. The predictive value of body protein for chemotherapy-induced toxicity. Cancer 2000;88: Lyman GH, Dale DC, Crawford J. Incidence and predictors of low dose-intensity in adjuvant breast cancer chemotherapy: a nationwide study of community practices. J Clin Oncol 2003;21: Savvides P, Terrin N, Erban J et al. Development and validation of a patient-specific predictive instrument for the need for dose reduction in chemotherapy for breast cancer: a potential decision aid for the use of myeloid growth factors. Support Care Cancer 2003;11: Szucs TD, Leonard R, Pettengell R et al. Dose-limiting effects of neutropenic events in six European audits of adjuvant breast cancer chemotherapy. Proc Am Soc Clin Oncol 2004;24: Silber JH, Fridman M, DiPaola RS et al. First-cycle blood counts and subsequent neutropenia, dose reduction, or delay in early-stage breast cancer therapy. J Clin Oncol 1998;16: Ray-Coquard I, Borg C, Bachelot T et al. Baseline and early lymphopenia predict for the risk of febrile neutropenia after chemotherapy. Br J Cancer 2003;88: Blay JY, Chauvin F, Le Cesne A et al. Early lymphopenia after cytotoxic chemotherapy as a risk factor for febrile neutropenia. J Clin Oncol 1996;14: 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 2000;18: Leong DC, Kinlay S, Ackland S et al. Low-risk febrile neutropenia in a medical oncology unit. Aust N Z J Med 1997;27: Kuderer NM, Cosler L, Crawford J et al. Cost and mortality associated with febrile neutropenia in adult cancer patients. Proc Am Soc Clin Oncol 2002;21:250a. 33 Kim YJ, Rubenstein EB, Rolston KV et al. Colony stimulating factors (CSFs) may reduce complications and death in solid tumor patients (Pts) with fever and neutropenia. Proc Am Soc Clin Oncol 2000;19:612a. 34 Gonzalez-Barca E, Fernandez-Sevilla A, Carratala J et al. Prognostic factors influencing mortality in cancer patients with neutropenia and bacteremia. Eur J Clin Microbiol Infect Dis 1999;18: Wilbur DW, Rentschler RE, Couperus JJ et al. Identifying neutropenic febrile cancer patients at risk for early death. Infect Med 2000;17: Viscoli C, Bruzzi P, Castagnola E et al. Factors associated with bacteraemia in febrile, granulocytopenic cancer patients. The International Antimicrobial Therapy Cooperative Group (IATCG) of the European Organization for Research and Treatment of Cancer (EORTC). Eur J Cancer 1994;30A: Rosenman M, Madsen K, Hui S et al. Modeling administrative outcomes in fever and neutropenia: clinical variables significantly influence length of stay and hospital charges. J Pediatr Hematol Oncol 2002;24: Kuderer NM, Dale D, Crawford J et al. The impact of febrile neutropenia in hospitalized cancer patients: morbidity, mortality, and cost. Proc Am Soc Clin Oncol 2004;23: Rackoff WR, Gonin R, Robinson C et al. Predicting the risk of bacteremia in children with fever and neutropenia. J Clin Oncol 1996;14: 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 1997;99: Santolaya ME, Alvarez AM, Becker A et al. Prospective, multicenter evaluation of risk factors associated with invasive bacterial infection in children with cancer, neutropenia, and fever. J Clin Oncol 2001;19: Klaassen RJ, Goodman TR, Pham B et al. Low-risk prediction rule for pediatric oncology patients presenting with fever and neutropenia. J Clin Oncol 2000;18:

11 Lyman, Lyman, Agboola et al Basu SK, Fernandez ID, Fisher SG et al. Study of mortality associated with febrile neutropenia among children with cancer. Proc Am Soc Clin Oncol 2004;24: Balducci L, Repetto L. Increased risk of myelotoxicity in elderly patients with non-hodgkin lymphoma. Cancer 2004;100: Balducci L, Extermann M. Cancer and aging. An evolving panorama. Hematol Oncol Clin North Am 2000;14: Repetto L, Biganzoli L, Koehne CH et al. EORTC Cancer in the Elderly Task Force guidelines for the use of colony-stimulating factors in elderly patients with cancer. Eur J Cancer 2003;39: National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines: Senior Adult Oncology v Available at Accessed May 5, Gomez H, Hidalgo M, Casanova L et al. Risk factors for treatment-related death in elderly patients with aggressive non-hodgkin s lymphoma: results of a multivariate analysis. J Clin Oncol 1998;16: Lyman GH, Crawford J, Dale DC et al. Predicting the risk of chemotherapy-induced neutropenia in patients with breast cancer: rationale for prospective risk model development. Poster presented at the 25th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 11 14, Lyman GH, Crawford J, Dale DC et al. Predicting the risk of neutropenic complications and reduced dose intensity in patients with early-stage breast cancer (ESBC): results from a prospective nationwide registry. Proc Am Soc Clin Oncol 2004;24: Dale DC, Wolff D, Agboola O et al. Prospective patient registry for the development of risk models of neutropenic complications including febrile neutropenia (FN). Blood 2002;100:502b. 52 Wolff DA, Crawford J, Dale DC et al. Risk of neutropenic complications based on a prospective nationwide registry of cancer patients initiating systematic chemotherapy. Proc Am Soc Clin Oncol 2004;24:547. ADDITIONAL READING Crawford J, Dale DC, Lyman GH. Chemotherapy-induced neutropenia: risks, consequences, and new directions for its management. Cancer 2004;100: Lyman GH. Risk assessment in oncology clinical practice. From risk factors to risk models. Oncology (Huntingt) 2003;17(suppl 11):8 13. Lyman GH, Dale DC, Crawford J. Incidence and predictors of low dose-intensity in adjuvant breast cancer chemotherapy: a nationwide study of community practices. J Clin Oncol 2003;21: Lyman GH, Dale DC, Friedberg J et al. Incidence and predictors of low chemotherapy dose-intensity in aggressive non- Hodgkin s lymphoma: a nationwide study. J Clin Oncol 2004;22: Lyman GH, Kuderer NM. The economics of the colony-stimulating factors in the prevention and treatment of febrile neutropenia. Crit Rev Oncol Hematol 2004;50: Lyman GH, Morrison VA, Dale DC et al. Risk of febrile neutropenia among patients with intermediate-grade non- Hodgkin s lymphoma receiving CHOP chemotherapy. Leuk Lymphoma 2003;44: