There has been a remarkable explosion in medicalinformation over the past several years.The rate of new discoveries and improved understandingof the biology and treatment ofcancer is ever-increasing. The same is true inthe area of supportive cancer therapy.[1]
There has been a remarkable explosion in medicalinformation over the past several years.The rate of new discoveries and improved understandingof the biology and treatment ofcancer is ever-increasing. The same is true inthe area of supportive cancer therapy.[1]
In the area of medical oncology we have beenconfronting the problem of neutropenia relatedto myelosuppressive chemotherapy treatmentfor decades. Prior to the 1990s, our only approachwas to educate chemotherapy patientsabout the risk of neutropenia, monitor theirblood counts periodically, and aggressivelymanage neutropenia-related fever and infection.[2,3] Because of the morbidity and mortalityassociated with these neutropenic events,patients who recovered adequately to undergofurther therapy generally received chemotherapydose reduction. This strategy was also usedas a primary approach at the initiation oftherapy for patients for whom the oncologistwas concerned that the risks and/or consequencesof development of neutropenia andinfection might outweigh the benefits of fulldosetreatment.
Treatment alternatives to change this paradigmfirst occurred in 1991, with the clinicalapproval of the colony-stimulating factors. Theuse of colony-stimulating factors as primaryprophylaxis after myelosuppressive chemotherapyallowed patients to reduce their risk ofneutropenic complications.[3] These agentsalso provided an opportunity to try to maintainfull-dose chemotherapy in patients who hadexperienced a neutropenic event on a priorchemotherapy cycle.
However, over the past decade, the strategy ofsecondary prophylaxis has been frequentlyemployed due to the limitations in our understandingof individual patient risk. Primaryprophylaxis has also been limited by the AmericanSociety of Clinical Oncology (ASCO)guidelines. Most studies demonstrating benefithave been performed in the setting of veryhigh-risk chemotherapy regimens where theexpected incidence of febrile neutropenia wasin excess of 40%. Since most chemotherapyregimens that are standard in clinical practicehave a substantially lower risk of febrile neutropenia,and because our patient risk modelshave not been well developed, the reactiveapproach for use of colony-stimulating factorshas been the norm.
The results of this "watch and wait" approachfor the supportive management of neutropeniaare clearly seen in several of the abstracts fromthe ASCO 40th Annual Meeting and the MultinationalAssociation for Supportive Care inCancer (MASCC)/International Society forOral Oncology (ISOO) 16th International Symposium,both held earlier this year. The frequencyof neutropenia in clinical practiceremains alarmingly high, and the complicationsin terms of inpatient hospitalization, morbidity,mortality, and economic cost from theUniversity HealthSystem Consortium databaseis staggering.
In 2002, a second-generation granulocytecolony-stimulating factor (pegfilgrastim[Neulasta]) was approved for use in the preventionof chemotherapy-related neutropenia.Clinical benefit from a single injectionpostchemotherapy was equivalent to that ofdaily dosing for more than 10 days of filgrastim(Neupogen) in randomized trials.[4] In additionto the advantages of improved technologyand better patient and provider convenience,subsequent clinical trials have not only helpedus better understand the utility of pegfilgrastim,but also helped us refine our use of colonystimulatingfactors.
Included in the 2004 ASCO/MASCC abstractsare important prospective studies that evaluatethe benefits of filgrastim or pegfilgrastim inprimary prevention of neutropenic complicationsin patient populations receiving chemotherapywhere the risk of febrile neutropenia issubstantially less than 40%. As the reader willsee in the pages ahead, risk reduction may beeven greater in these lower-risk myelosuppressivesettings.
In addition to improving our understanding ofthe epidemiology of neutropenia and its consequencesand the ability to redefine our thresholdof primary prophylaxis with colony-stimulating factors, the 2004 ASCO andMASCC/ISOO abstracts also provide a wealthof information in the disease-specific areas ofbreast cancer, lung cancer, lymphoma, andhematologic malignancies. I invite the readerto review all of the sections carefully to betterrefine your understanding of the preventionand management of neutropenia in cancer chemotherapypatients within your own practice.In 2005 the ASCO guidelines will be updated,and we will see for the first time the NationalComprehensive Cancer Network guidelines onthe use of colony-stimulating factors. The impactof the abstracts included in this supplementto ONCOLOGY will be substantial. As aresult, we will hopefully move closer to reducingthe morbidity of chemotherapy while improvingour ability to deliver full-dose therapywhen appropriate to enhance overall patientoutcomes. For this to happen, the oncologycommunity must work closely together not onlyto conduct the appropriate research, but toeducate one another about the results. I amhopeful that the abstracts and commentary inthe pages ahead will help us all in that process.
Assessing the Risks andConsequences ofNeutropenia
Abstract #6125
Risk of Neutropenic ComplicationsBased on a Prospective NationwideRegistry of Cancer Patients InitiatingSystematic Chemotherapy
D. A. Wolff, J. Crawford, D. C. Dale, M. S. Poniewierski,G. H. Lyman, for the ANC Study GroupUniversity of Rochester Medical Center, Rochester, New York;Duke University Medical Center, Durham, North Carolina;University of Washington Medical Center, Seattle, Washington
Myelosuppression represents the major dose-limitingtoxicity of cancer chemotherapy. A prospective,nationwide study was undertaken to define risk factorsfor neutropenic complications (NC) and to developrisk models for selecting patients for hematopoieticsupport. More than 2,500 patients have been prospectivelyregistered at 137 randomly selected practicesites.Primary outcomes included documented severeneutropenia (SN) (ANC < 500), NC (fever/infection/other), and severe neutropenic events (SNE) (SN orNC). This preliminary analysis is based on the first2,222 patients treated to date including 886 (40%)age ≥ 65. One-third did not receive at least four cyclesof chemotherapy due to disease progression (55%),refusal (11%), death (11%), and unknown (23%).Neutropenia (ANC < 1,000) was documented in40%, including SN in 26%. More than one-half of allinitial events occurred during cycle 1.In addition to cancer and regimen type, significantpredictors of SN included gender (.001), baselineneutrophils (.01), diabetes (.037), and chronic lungdisease (.001). Significant predictors of NC includedgender (.004) and stage (.04). In addition to cancerand regimen type, significant predictors of SNE includedgender (.006), stage (.007), and chronic obstructivepulmonary disease (.029). The risk of SNEin all cycles (first cycle) includes breast 49 (36), lung31 (20), lymphoma 42 (33), ovary 37 (17), and colon18 (8). See Table 1 for summary.
CONCLUSION: Neutropenic complications andsevere neutropenic events occur early in the course oftherapy, potentially compromising chemotherapy doseintensity and clinical outcomes. Significant independentrisk factors include cancer type, regimen, doseintensity, gender, stage, and several comorbidities.
Abstract #A-29
Assessment of Neutropenic Riskin Cancer Patients ReceivingSystemic Chemotherapy:Results From a ProspectiveNationwide Registry
J. Crawford, D. Wolff, D. C. Dale, E. Culakova,M. Poniewierski, G. H. Lyman, for the ANC Study GroupDuke University, Durham, North Carolina; University ofRochester, Rochester, New York; University of Washington,Seattle, Washington
Myelosuppression, including chemotherapy-inducedneutropenia (CIN), is the major dose-limitingtoxicity of cancer chemotherapy. This prospective,nationwide study was undertaken to better define riskfactors for CIN in treatment-naive or previouslytreated patients with five major tumor types (breast,lung, colon, ovarian, and lymphoma). More than2,500 patients initiating a chemotherapy regimenhave been prospectively registered at 137 randomlyselected practice sites. Primary outcomes includesevere neutropenia (SN; ANC < 500), neutropeniccomplications (NC; fever/infection/other), and severeneutropenic events (SNE; SN or NC).Of the first 2,222 patients enrolled, 886 (40%)were > 65 years of age. One-third of patients receivedfewer than four cycles of chemotherapy due to diseaseprogression (55%), refusal (11%), death (11%),or other unknown reasons (23%). Neutropenia (ANC< 1,000) and severe neutropenia were documented in40% and 26% of patients, respectively. Neutropeniccomplications occurred early in the course of treatmentin all tumor types, with 67% of all SNE in cycle1. The risk of SNE (all cycles; first cycle) variedgreatly with disease: breast (49%; 36%), lung (31%;20%), lymphoma (42%; 33%), ovary (37%; 17%),colon (18%; 8%).In addition to tumor type and treatment regimen,significant predictors (P value) of SN included gender(.001), baseline neutrophils (.01), and comorbiditiesincluding diabetes (.037) and chronic lung disease(.001). Significant predictors of NC included femalegender (.004) and stage (.04) with borderline significancefor previous neutropenia and comorbidities. Inaddition to cancer and regimen type, significant predictorsof SNE included female gender (.006), stage(.007), and chronic lung disease (.029), with borderlinesignificance for baseline white blood cell andneutrophil count.CONCLUSION: This prospective registry holdspromise for the development of a reliable, generalizedrisk model that will allow the accurate predictionof neutropenic risk in individual patients, enablingthe selection of patients who may benefit from prophylactichematopoietic support.
Abstract #6049
Complications and Costs AssociatedWith Febrile Neutropenia inHospitalized Adult Cancer Patients
N. M. Kuderer, J. Crawford, D. C. Dale, G. H. Lyman, for theANC Study GroupUniversity of Rochester Medical Center, Rochester, New York;Duke University Medical Center, Durham, North Carolina;University of Washington Medical Center, Seattle, Washington
Fever in the setting of neutropenia (FN) frequentlyrequires hospitalization for empiric broad-spectrumantibiotics. More complete understanding of themedical complications and costs associated with hospitalizationfor FN is needed.Hospitalization of cancer patients with FN at 115academic medical centers between 1995 and 2000was studied using the discharge database of the UniversityHealthSystem Consortium. Primary outcomesincluded length of stay (LOS), cost, infectious complications,and mortality. Logistic regression analysiswas used to estimate relative risk based on adjustedodds ratios (OR). Hospitalization with FN was reportedin 41,779 adult cancer patients. Mean (median)LOS was 11.2 (6) days while the average (median)cost was $19,110 ($8,376) per episode. The 35% ofpatients hospitalized ≥ 10 days accounted for 74% ofhospital days and 78% of total cost. Over the 6 yearsof observation, LOS decreased 10% while cost perday and total cost increased 28% and 13%, respectively.Documented infection was reported in 38% ofpatients including sepsis in 19% and pneumonia in10%. Death during hospitalization was reported in11% of admissions. Mortality was most commonlyassociated with gram-negative sepsis (34%), pneumonia(27%), or comorbidities including renal (30%),cerebrovascular (30%), liver (28%), and lung disease(27%). Mortality rates increased with the number ofcomorbidities (
P
trend < .001).Odds ratios for death were as follows: gramnegativesepsis (4.8), pneumonia (2.4), renal (3.1),cerebrovascular (3.2), liver (2.9), and lung disease(3.9). Odds ratios for complicated hospitalization(length of stay ≥ 10 days) were leukemia (3.4), grampositivesepsis (2.5), and comorbidities with lung(2.1), renal (2.1), cerebrovascular (2.1), and liverdisease (2.0) after adjustment for other comorbiditiesand infectious complications.CONCLUSION: Hospitalization for FN is associatedwith considerable morbidity and mortality. Numerouspatient characteristics, comorbidities, andinfectious complications are associated with increasedmortality, length of stay, and cost.
Abstract #A-57
Adult Cancer Patients HospitalizedWith Febrile Neutropenia: RiskStratification Based on an Analysisof the University HealthSystemConsortium Discharge Database
N. M. Kuderer, J. Crawford, D. C. Dale, G. H. Lyman for theANC Study GroupUniversity of Rochester, Rochester, New York; Duke University,Durham, North Carolina; University of Washington, Seattle,Washington
Febrile neutropenia (FN), the major dose-limitingtoxicity of cancer chemotherapy, frequently requireshospitalization for the administration of empiric broadspectrum antibiotics. For a better understanding ofthe clinical impact and economics of hospitalizationfor FN, the records of 41,779 adult nontransplantcancer patients admitted with FN at 115 academicmedical centers between 1995 and 2000 were analyzedusing the discharge database of the UniversityHealthSystem Consortium.Primary outcomes included length of stay (LOS),cost, infectious complications, and mortality. Averageage was 53.6 years with 28% ≥ 65 years. Mean(median) LOS was 11.2 (6) days while the average(median) cost was $19,110 ($8,376) per episode. The35% of patients hospitalized for ≥ 10 days accountedfor 74% of hospital days and 78% of total cost.Documented infection and death during hospitalizationwere reported in 38% of patients and 11% ofadmissions, respectively. Mortality rates increasedwith the number of comorbidities (
P
-trend < .0001).In multivariate analysis, risk factors (odds ratios)for inpatient mortality were gram-negative sepsis(4.84), pneumonia (2.33), and renal (3.19), cerebrovascular(3.29), liver (2.93), and lung disease(3.95); predictors for LOS ≥ 10 days were grampositivesepsis (2.39), leukemia (3.54), and variouscomorbidities. Risk scores for LOS ≥ 10 days (range:0-32) and inpatient mortality (range: 0-34) werederived from the logistic regression models. Theserisk scores effectively discriminated patients at highrisk for prolonged hospitalization with FN and inpatientmortality. The 24% of patients with risk scores> 4 experienced a risk of ≥ 50% for LOS ≥ 10 days.Similarly, the 27% of patients with risk scores > 5experienced a risk of ≥ 12% for inpatient mortality.CONCLUSION: By identifying individual cancerpatients at increased risk for prolonged hospitalizationand inpatient mortality, the risk score analysismay help guide supportive care treatment decisions,improve outcomes, and potentially reduce the cost ofcancer care.
Abstract #6060
Human Resource Costs and PatientTime Affected By the Delivery ofChemotherapy and NeutropeniaManagement
B. V. Fortner, T. A. Okon, L. Zhu, K. Tauer, K. Moore,D. TempletonSupportive Oncology Services, Memphis, Tennessee;The West Clinic, Memphis, Tennessee
The purpose of this study was to evaluate humanresource costs and patient time associated with thedelivery of chemotherapy and management of chemotherapy-induced neutropenia (CIN).A total of 400 medical professionals were surveyedregarding human resource time associatedwith medical tasks, and 189 patients were surveyedregarding time and activities affected by medicalvisits across 20 community oncology practices.Results [mean (standard deviation)] showed chemotherapyand CIN-related medical visits involvenumerous types of professionals (X = 10 per practice)who execute multiple medical tasks (X = 230 perpractice), resulting in substantial human resourcetime and expense to the practice. For example, 1 dayof chemotherapy X = 4.23 (1.48) hours, $152.55($65.89); midcycle lab visit X = 2.09 (1.05) hours,$48.62 ($28.22); 5 days of IV antibiotics X = 15.7(6.1) hours, $415.9 ($213.6); 10 days of filgrastim X= 24.4 (11.1) hours, $579.30 ($292.60); 1 day ofpegfilgrastim X = 2.40 (1.12) hours, $57.06 ($30.94).Furthermore, results showed that even relatively simplemedical visits resulted in large disruptions ofpatient time and life activities before, during, andafter the visit. For example, 1 day of chemotherapy X= 8.19 (3.93) hours; midcycle lab visit X = 2.27 (0.92)hours; 5 days of IV antibiotics X = 16.31 (4.92) hours;10 days of filgrastim X = 23.2 (9.01) hours; 1 day ofpegfilgrastim X = 2.36 (1.44) hours.Sensitivity analysis demonstrated that as rates ofsevere CIN increase, human resource costs and patientburden increase. At relatively high rates of CIN,prophylactic use of growth factors was predicted tobe less burdensome to patients and more cost-effectivein terms of human resource costs.CONCLUSION: These data are important for understandingthe cost implications of delivering chemotherapyin the community oncology setting. Thesedata are also important for understanding the burdenmultiple medical visits place on patients and theircaregivers and may help guide the creation of supportivecare strategies that minimize unnecessarypatient and caregiver burden.
Commentary on Abstracts #6125,#A-29, #6049, #A-57, and #6060
Jeffrey Crawford, MD
To understand the magnitude of complicationsof chemotherapy, in the past we have relied on theresults of clinical trials. Unfortunately, results ofthese trials led us to underappreciate the severityand magnitude of neutropenia for several reasons.First and most importantly, the clinicaltrials were done in a selected population of patientsof generally better performance status,younger age, and with less prior treatment, andtherefore were often not comparable to clinicalpractice. In addition, even phase III trials wereuneven with reference to laboratory monitoringand reporting of neutropenic events. Therefore,the best way to understand the true magnitude ofneutropenia and to identify patients at risk isthrough population-based databases.
Perhaps one of the most important studies toassess the magnitude of neutropenia in communitypractice is the ongoing prospective nationwideregistry of cancer patients initiating systemicchemotherapy as reported by the ANC study group(ASCO abstract #6125 and MASCC abstract#A-29). At the times of these reports more than2,500 patients had been prospectively registeredfrom 137 randomly selected practice sites. Fortypercent of the population is over the age of 65.Neutropenia occurred in 40% of the patients andsevere neutropenia in 26%. More than half of allthe initial neutropenic events occurred in the firstcycle of treatment, speaking to the problem of the"watch and wait" approach. At the time of thisanalysis, risk factors for severe neutropenia variedby tumor type and treatment regimen, but alsoincluded female gender, stage, and comorbidity,particularly lung disease. As this registry is completed,it is hoped that a prospective risk modelcan be developed for validation in subsequenttrials.
Equally impressive from the ANC study groupreported by Kuderer and colleagues (ASCO abstract#6049, MASCC abstract #A-57) are theresults of a retrospective analysis of the UniversityHealthSystem Consortium (UHC) dischargedatabase. This administrative database providedrecords on more than 40,000 adult nontransplantcancer patients who were admitted with febrileneutropenia at 115 academic medical centersbetween 1995 and 2000. While the median lengthof stay for febrile neutropenia was 6 days, themean number of days was 11, and 35% of patientswere hospitalized for 10 days or longer. This 35%of patients accounted for 78% of the total cost.Documented infections occurred in 38% of patientsand 11% of patients died during hospitalization-alarmingly high numbers. Mortality rateswere clearly related to the number of comorbiditiesas outlined in the abstract. Better understandingof the population of patients who notonly have experienced febrile neutropenia, buthave experienced the worst complications, mayalso help refine prospective risk models.
Another interesting abstract presented at ASCOby Fortner and colleagues (ASCO abstract #6060)looked at human resources and patient time affectedby neutropenia management. Even in theabsence of severe complications such as hospitalization,the time spent for simple medical visitsand disruption of patient time and life activitieswas substantial for the patient and caregiver.Equally important was the impact on the oncologist'soffice practice. Such analyses are importantif we are to truly measure the economic andpersonal costs involved with neutropenia management.
Redefining the Febrile Neutropenia Risk Threshold
Abstract #8002
Prevention of Chemotherapy-InducedFebrile Neutropenia by Antibiotics vsAntibiotics Plus Granulocyte Colony-Stimulating Factor in Small-Cell LungCancer: A Randomized Phase III Study
J. Timmer-Bonte, B. Biesma, J. Smit, F. Wilschut,J. Akkermans, T. De Boo, G. Bootsma, V. C. Tjan-HeijnenUniversity Medical Centre Nijmegen, Nijmegen, Netherlands;Jeroen Bosch Hospital, Den Bosch, Netherlands; RijnstateHospital, Arnhem, Netherlands; Gelderse Vallei Hospital, Ede(Gld), NetherlandsSupported by a research grant from the Dutch Healthcare InsuranceBoard
Febrile neutropenia (FN) is a major complicationof chemotherapy in patients with small-cell lungcancer (SCLC). Granulocyte colony-stimulating factor(G-CSF) is used to prevent FN, but its primary useis only recommended in patients with a considerable(> 40%) risk of FN (ASCO guidelines,
J Clin Oncol
2000). Antibiotics are also effective in preventingFN, even reduce infection-related mortality, and arecost-effective (Tjan-Heijnen VC:
Ann Oncol 2001
).In this multicenter randomized phase III trial, therole of combined primary prophylaxis with antibioticsplus G-CSF in patients with SCLC at increasedrisk of FN was determined. Patients with SCLClimiteddisease and performance status (PS) 2/3, orage over 60 years or judged not suitable for concurrentchemo/radiotherapy, and patients with SCLCextensivedisease were considered at increased risk ofFN. Patients were stratified for age (< / ≥ 60 years), PS(0-1/2), extent of disease (limited/extensive), andline of therapy (first/second), and randomized forprimary prophylaxis with antibiotics (ciprofloxacin500 mg + roxithromycin 150 mg, bid, days 4-13)alone or in combination with G-CSF (filgrastim 5 μg/kg/d, days 4-13). Chemotherapy consisted of cyclophosphamide1,000 mg/m
2
day 1, doxorubicin 45 mg/m
2
day 1, and etoposide 100 mg/m
2
days 1, 2, 3, IVq3wk, * 5.The primary end point was incidence of FN in firstcycle. Targeted accrual was N = 156 patients (78/arm).See Table 1 for results. A total of 171 patients wereeligible. Patient characteristics were well balanced.
CONCLUSION: The addition of primary prophylacticG-CSF to prophylactic antibiotics significantlyreduced the incidence of chemotherapy-related FN inSCLC patients at increased risk of FN, especially in thefirst cycle by 50%.
Abstract #620
Prophylactic Growth FactorSupport With Adjuvant Docetaxel,Doxorubicin, and Cyclophosphamidefor Node-Negative Breast Cancer:An Interim Safety Analysis of theGEICAM 9805 Study
M. Martin, A. Lluch, M. A. Segu, A. Antn, A. Ruiz,M. Ramos, A. Rodrguez-Lescure, E. AdroverSpanish Breast Cancer Research Group (GEICAM), Madrid,Spain
For patients with node-positive breast cancer, TAC(docetaxel, doxorubicin, and cyclophosphamide)confers significant disease-free and overall survivalbenefits vs FAC (fluorouracil, doxorubicin, cyclophosphamide),but with a higher rate of febrile neutropenia(Martin M: SABCS 2003, abstract 43). Inour study of TAC vs FAC for node-negative breastcancer, we performed an interim safety analysis toassess the impact of growth factor support on theincidence of TAC-related adverse events.Following surgery, patients with operable, highrisk(St Gallen, 1998), node-negative breast cancer,18-70 years old, Karnofsky performance status ≥80%, and adequate hematologic and organ functionwere randomized to FAC (F 500 mg/m
2
, A 50 mg/m
2
,C 500 mg/m
2
) or TAC (T 75 mg/m
2
, A 50 mg/m
2
, C500 mg/m
2
) day 1 q3wk for six cycles. After enrollmentof 224 patients, the study was amended torequire prophylactic G-CSF for patients subsequentlytreated with TAC, but not FAC. The presentanalysis assessed the impact of G-CSF on the incidenceof febrile neutropenia (FN; fever ≥ grade 2 withgrade 4 neutropenia) and other grade 3/4 toxicities inpatients treated with TAC.At the cutoff date for this analysis, 448 patients hadbeen enrolled: 124 received FAC (111 prior to amendment)and 124 TAC (109 without mandatory G-CSF[TAC-G]; 115 with G-CSF [TAC+G]). For patientsreceiving FAC, the incidence of FN (% patients) was1.3% (0.9% pre-, 1.7% postamendment); other grade3/4 adverse events were observed in 26.7% (27% preand26.5% postamendment). Among patients treatedwith TAC, the rates of febrile neutropenia were23.8% for TAC-G (95% CI = 15.9%-31.9%) and3.5% for TAC+G (1.0%-8.7%); the rates of othergrade 3/4 toxicities were 50.4% for TAC-G (41.1%-59.8%) and 20% for TAC+G (12.7%-27.3%). Therelative dose intensities for TAC-G vs TAC+G, respectively,were T: 92% vs 96%, A: 93% vs 97%, C:93% vs 97%.CONCLUSION: Within the limitations of a nonrandomizedcomparison, the use of G-CSF beginningwith the first cycle of TAC substantially reduces theincidence of FN and other grade 3/4 toxicities andenables maintenance of chemotherapy dose intensityfor women with early stage breast cancer. The adverseevent rate with TAC + G is similar to that ofFAC.
Abstract #A-52
Prophylactic PegfilgrastimSignificantly Reduces the Incidenceof Febrile Neutropenia,Hospitalizations, and IV Anti-InfectiveUse in Patients With Breast CancerReceiving Docetaxel: A Phase III,Randomized, Double-Blind,Placebo-Controlled Study
L. Schwartzberg, S. Tjulandin, M. Wljtukiewicz,L. J. Barajas-Figueroa, M. Palmer, T. NeumannWest Cancer Clinic, Memphis, Tennessee; Blokhin CancerResearch Center, Moscow, Russia; Regional Oncology Centre,Bialystok, Poland; Hospital General de Occidente, Zapopan,Jalisco, Mexico; Amgen Inc, Thousand Oaks, California
Previous studies have demonstrated that once-percyclepegfilgrastim (Neulasta) is as effective as dailyfilgrastim (Neupogen) for the reduction of febrileneutropenia (FN) in patients with breast cancer receivingchemotherapy regimens associated with ahigh risk of FN. Current treatment guidelines recommendprophylactic use of colony-stimulating factorsfor treatment regimens associated with a ≥ 40%incidence of FN.To test whether patients at moderate risk of FNmay also benefit from once-per-cycle. first-cycle useof pegfilgrastim, this phase III randomized, doubleblind,placebo-controlled study evaluated the effectsof prophylactic, fixed doses of pegfilgrastim on theincidence of FN in patients with breast cancer receivingdocetaxel; which is associated with an averagereported FN incidence of 20% in the absence ofgrowth factor support. Patients with breast cancer(stage II to IV) and an ECOG performance status of0 to 2 who were candidates for docetaxel chemotherapy(100 mg/m
2
every 3 weeks) were eligible forparticipation.Patients were randomized in a 1:1 allocation toreceive either 6 mg of pegfilgrastim or placebo onceper cycle on the day after docetaxel administrationfor up to four cycles. Febrile neutropenia was definedas a temperature ≥ 38.2
o
C and absolute neutrophilcount (ANC) < 0.5 * 10
9
/L (measured on the same dayor the day after a temperature ≥ 38.2
o
C). A total of928 patients were randomized to receive pegfilgrastim(n = 463) or placebo (n = 465). The percentage ofpatients developing FN was statistically significantlylower in the pegfilgrastim group (1% [6/463]) compared with the placebo group (17% [78/465];
P
< .0001). Pegfilgrastim was also associated with asignificantly lower incidence of hospitalizations (1%[6/463] vs 14% [64/465];
P
< .0001) and intravenousanti-infective use (2% [7/463] vs 10% [48/465]:
P
< .0001) compared with the placebo group.CONCLUSION: Pegfilgrastim was well toleratedand significantly decreased the incidence of FN,hospitalizations, and anti-infective use when usedprophylactically.
Commentary on Abstracts #8002,
#620, and #A-52
Jeffrey Crawford, MD
Given the magnitude of neutropenia and itsconsequences in the cancer chemotherapy patient,preventive strategies would certainly appearto be warranted. As mentioned in theintroduction, for nearly a decade the ASCO guidelineshave recommended that primary prophylaxiswith colony-stimulating factors be restricted topatients who are at a 40% or greater risk for thedevelopment of febrile neutropenia. This wasbased on the control arms of the initial randomizedclinical trials and was also supported byinitial pharmacoeconomic data. More recent pharmacoeconomicdata that take into account therising cost of hospitalization and other expenses([Lyman G, et al: Eur J Cancer 34[12]:1857-1864, 1998) have suggested that the risk thresholdshould be closer to 20% for the use ofcolony-stimulating factors to be relatively costneutral.However, this economic data precededclinical trial data supporting the efficacy of colony-stimulating factors in this range, until thedata from ASCO and MASCC of 2004.
At ASCO, Timmer-Bonte and colleagues (ASCOabstract #8002) performed a prospective randomized-based study to evaluate the impact offilgrastim on patients with small-cell lung cancerreceiving cyclophosphamide, doxorubicin, andetoposide, along with prophylactic antibioticsincluding ciprofloxacin and roxithromycin. Forthe group who received antibiotics alone, theincidence of febrile neutropenia was 23% in thefirst cycle and 30% in overall cycles. For thegroup receiving G-CSF, the incidence of febrileneutropenia was 10% in the first cycle and 18%overall, both statistically significant. While theabsolute numbers for mortality from febrile neutropeniawere low, the trend was in the samedirection, with 7% of patients on the antibioticarm dying compared to 3% of patients on theantibiotic plus G-CSF arm.
A second study performed by the Spanish breastcancer research group, reported by Martin andcolleagues (ASCO abstract #620), looked at prophylacticgrowth factor support in patients receivingadjuvant docetaxel, doxorubicin, andcyclophosphamide (TAC). In this study of womenwith node-positive breast cancer, the incidence offebrile neutropenia was high with the TAC regimenprompting an amendment in this study toinclude prophylactic G-CSF. The rates of febrileneutropenia were 23.8% prior to the institution ofG-CSF and 3.5% for the patients enrolled in thesubsequent part of the study. While there arelimitations from this nonrandomized comparison,this magnitude of difference strongly suggeststhat G-CSF was effective in reducing therisk of febrile neutropenia with this regimen.
The most convincing study was reported bySchwartzberg at MASCC (abstract #A52). In thistrial, more than 950 breast cancer patients wereenrolled in a prospective, double-blind placebocontrolledtrial evaluating the impact of a singledose of pegfilgrastim 24 hours after docetaxel at100 mg/m2. This regimen was specifically chosento try to assess the potential benefits of growthfactor support in the setting associated with approximately20% risk of febrile neutropenia.The results were striking. The placebo-controlgroup experienced a 17% incidence of febrileneutropenia compared to a 1% incidence in thepegfilgrastim group. This corresponded with reductionin hospitalization from 14% to 1%, andreduction in intravenous antibiotic use from 10%to 2% in favor of the pegfilgrastim group.The results of these three trials are strikinglyconsistent and suggest a benefit at least as greatas or greater than the previous clinical trials thatevaluated G-CSF in treatment settings where therisk of febrile neutropenia was higher. In view ofthese benefits, it is anticipated that the guidelinescommittee of ASCO and NCCN will reevaluatethe paradigm for treatment recommendations forprimary prophylaxis with colony-stimulating factorsin the cancer chemotherapy patient. As wedefine the patient at risk not only by the diseaseand chemotherapy regimen, we must also defineindividual factors that may place a patient at riskof febrile neutropenia. The appropriate interventionscan then be made to reduce the currentmagnitude of and consequences from chemotherapy-induced neutropenia.Understanding the Duration of TherapyAbstract #A-50A Meta-Analysis of Filgrastim DosingDuration for the Treatment ofNeutropenia From Standard andDose-Intensified ChemotherapyG. D. Demetri, Y. Mun, B. McGuire, R. D. BaynesDana-Farber Cancer Institute, Boston, Massachusetts; AmgenInc, Thousand Oaks, CaliforniaA meta-analysis was conducted to study the averagenumber of days of filgrastim dosing within a cycle thatwere needed to achieve neutrophil (ANC) recoveryfollowing myelosuppressive chemotherapy. Aggregatestatistics from 10 phase I or phase II studies were used(see Table 1). Each study treated a specific solid orlymphoid tumor with conventional chemotherapy regimens,first at standard doses and then at incrementallyintensified doses using dose escalation and/or shortenedcycle lengths (dose density).
Filgrastim was dosed at 5 or 10 μg/kg/d startingafter chemotherapy and through two ANC observations≥ 10 * 109 L. Data were available from 334patients: 138 who received standard chemotherapydoses and 196 who received escalated chemotherapy.Total days of filgrastim administered in cycle 1 wereanalyzed using a random effects meta-analysis model.For cohorts receiving standard chemotherapy (includingshortened cycles), the pooled mean days offilgrastim was 12.4 days (95% CI = 9.8-15.0 days).Individual study means ranged from 2.5 to 17.0 days.The lowest mean value (2.5) represented a studyusing a nitrosourea; which is known to elicit mild anddelayed neutropenia. In patients receiving escalateddoses of chemotherapy, the pooled mean of filgrastimdosing duration in cycle 1 was 15.6 days (95% CI =13.7-17.4 days). At high doses of chemotherapy,delayed initiation of filgrastim or premature discontinuationwas associated with prolonged neutropenia(or failure to achieve a sustained ANC recovery) andsubsequent cycle delays.CONCLUSION: This analysis demonstrated thatin the setting of standard doses of chemotherapy,approximately 12 days per cycle of filgrastim wererequired to achieve this level of ANC recovery. Whenchemotherapy was escalated, longer durations ofdosing were required. Initiating filgrastim dosing toolate in the cycle and/or stopping too soon were foundto be deleterious to achieving a stable ANC fordelivery of the next cycle of chemotherapy.Abstract #A-59Duration of G-CSF Prophylaxisand Risk of Hospitalization AmongPatients With Non-Hodgkin'sLymphoma, Breast Cancer,and Lung CancerD. Weycker, J. Hackett, J. Edelsberg, G. Oster, A. GlassPolicy Analysis Inc, Brookline, Massachusetts; Amgen Inc,Thousand Oaks, California; Kaiser Permanente Northwest,Portland, OregonIn clinical trials, granulocyte colony-stimulatingfactor (G-CSF) prophylaxis, when administered foran average of 10 to 11 days, has been found to reducethe incidence of febrile neutropenia in patients receivingmyelosuppressive chemotherapy. In clinicalpractice, however, many patients receive shortercourses of prophylaxis. The effectiveness of theseshorter courses is unknown.Using a large US health-care claims database, weidentified all adults with non-Hodgkin's lymphoma(NHL), breast cancer (BC), or lung cancer (LC) whoreceived myelosuppressive chemotherapy between1998 and 2002. For these persons, we further identifiedtheir first such course of chemotherapy and eachunique cycle within that course. We then focusedattention on all patient-cycles in which G-CSF wasadministered on or before cycle day 5 ("G-CSFprophylaxis"). Pooling all such cycles, we used aGeneralized Estimating Equation (GEE) model (witha logistic link function) to examine the relationshipbetween duration of G-CSF prophylaxis and risk ofhospitalization for neutropenia or infection and riskof hospitalization for any reason, controlling forpotential confounders.Mean (SD) duration of G-CSF prophylaxis was6.5 (3.1) days across 332 patient-cycles in those withNHL, 6.1 (2.9) days across 482 patient-cycles inthose with BC, and 4.3 (3.1) days across 522 patientcyclesin those with LC. There were 29 cycles with ahospitalization for neutropenia or infection amongNHL patients, 21 among BC patients, and 45 amongLC patients. The predicted risk of hospitalization forneutropenia or infection declined with each additionalday of G-CSF prophylaxis for patients with NHL(odds ratio [OR] = 0.81; P = .003), BC (OR = 0.78; P= .002), and LC (OR = 0.92; P = .119) (see Table 1).Results were similar for analyses of hospitalizationfor any reason.
CONCLUSION: Among patients with NHL, BC,and LC receiving G-CSF prophylaxis, longer coursesof therapy confer lower risks of hospitalization.Commentary on Abstracts #A-50and #A-59
Jeffrey Crawford, MDAt MASCC, two studies evaluated the impact ofthe duration of filgrastim dosing on the risk ofdevelopment of neutropenia and its consequences.In the report from Demetri and colleagues(abstract #A-50), a meta-analysis was performedthat demonstrated that in the setting of standarddosechemotherapy, approximately 12 days percycle of filgrastim was required to achieve adequateANC recovery. With escalated doses ofchemotherapy even longer durations of dosingwere required.Initiating G-CSF too late in the cycle or stoppingtoo soon were also found to be less beneficialin achieving stable ANC. In abstract #A-59, Weyckerand colleagues report results of G-CSFprophylaxis using a large US health-care claimsdatabase. Interestingly, in this very different dataset, the results were similar to the meta-analysisof clinical trials reported by Demetri. That is,among patients with non-Hodgkin's lymphoma,breast cancer, and lung cancer who receivedG-CSF prophylaxis, longer courses of therapyconfer lower risks of hospitalization.These results are quite important to help validatethe necessity of prolonged use of G-CSF withmyelosuppressive chemotherapy regimens to reducethe risk of neutropenia. Since the registrationaltrials for pegfilgrastim included a controlarm of filgrastim use of 11 days, this also validatesthe appropriateness of current pegfilgrastimdosing in order to achieve full benefits of eitheragent in the setting of myelosuppressive chemotherapy.Neutropenia and Lung CancerAbstract #7223Risk and Mortality AssociatedWith Febrile Neutropenia inLung Cancer PatientsD. B. Daniel, J. Crawford, N. M. Kuderer, D. C. Dale,G. H. Lyman, for the ANC Study GroupDuke University Medical Center, Durham, North Carolina;University of Rochester Medical Center,Rochester, New York; University of Washington MedicalCenter, Seattle, WashingtonMyelosuppression and its complications representthe major dose-limiting toxicities of lung cancerchemotherapy. Length of stay (LOS), mortality, andcost associated with febrile neutropenia were studiedin 3,340 lung cancer patients experiencing 3,846admissions to 115 academic institutions reporting tothe University HealthSystem Consortium between1995 and 2000. Those factors found to be significantlyassociated with complicated hospitalizations (lengthof stay ≥ 8 days) or inpatient mortality were evaluatedin logistic regression models. Estimates of relativerisk associated with each risk factor were based on theadjusted odds ratios (OR ± 95% CIs) for mortalityand LOS greater than 8 days.Overall inpatient mortality was 12.1%. Factorssignificantly associated with increases in the risk ofinpatient mortality included metastases (OR 2.28[95% CI = 1.81-2.88]), hypotension (OR 3.79 [95%CI = 2.55-5.62]), pneumonia (OR 2.39 [95% CI =1.90-3.01]), and gram-negative bacteremia (OR age≥ 65: 8.13; OR age ≤ 65: 3.75). The mean and medianLOS were 8.1 and 5 days, respectively. The meanLOS was 8.1 days, median 5 days. Hospital LOS was≥ 8 days in 32.4% patients. Factors significantlyassociated with longer LOS included congestive heartfailure (OR 2.18 [95% CI = 1.61-2.94]), functionalimpairment (OR 1.98 [95% CI = 1.35-2.90]), pneumonia(OR 1.94 [95% CI = 1.63-2.31]), and grampositivebacteremia (OR 3.42 [95% CI = 2.30-5.11]).Patients whose primary reason for admission wasother than febrile neutropenia had higher risk ofmortality (OR 3.40 [95% CI = 2.49-4.64]) and LOS≥ 8 days (OR 3.18 [95% CI = 2.68-3.77]). Model R2for the mortality and LOS models were 0.126 and0.123, respectively, while the c-statistics (ROC AUC)were .790 and .718.CONCLUSION: Mortality in lung cancer patientshospitalized for febrile neutropenia is greater than forother solid tumors. Increasing age, advanced disease,nutritional and functional impairments, and comorbiditiesalong with infectious complications lead to prolongedhospitalization and increased risk of mortality.Prospective studies are warranted to further define riskfactors and evaluate possible interventions.Abstract #7007Japan-SWOG Common Arm Analysisof Paclitaxel/Carboplatin in AdvancedStage Non-Small-Cell Lung Cancer:A Model for Prospective Comparison ofCooperative Group TrialsD. R. Gandara, Y. Ohe, K. Kubota, Y. Nishiwaki, Y. Ariyoshi,N. Saijo, S. Williamson, P. N. Lara, J. Crowley, M. FukuokaUniversity of California Davis Cancer Center, Sacramento,California; FACS Cooperative Group, Tokyo, Japan; Universityof Kansas, Kansas City, Kansas; Southwest OncologyGroup, San Antonio, TexasWhether results of clinical trials performed outsidethe United States can be fully extrapolated to USpopulations remains in question due to potentialdifferences in trial designs, study-specific criteria,patient demographics, and population-related pharmacogenomics.We prospectively designed and conducted separatephase III trials in advanced non-small-cell lungcancer (NSCLC) linked by a "common arm" withidentical eligibility, staging, response, and toxicitycriteria, to (1) determine similarities and differencesin patient demographics and outcomes in cooperativegroup trials in Japan (Four-Arm Cooperative Studyor FACS) and the United States (SWOG 0003), (2)provide the basis for global standardization in clinicaltrials in NSCLC, and (3) facilitate regulatory changesneeded for joint Japan-US studies sponsored by theUS National Cancer Institute (NCI). We performed aplanned comparative analysis of the paclitaxel/carboplatinarms from FACS and S0003, identical exceptfor paclitaxel dose of 200 mg/m2 in FACS and225 mg/m2 in S0003, based on the MTD from separatephase I studies in Japan and the US. CarboplatinAUC 6 was used in both. See Table 1:
CONCLUSION: (1) This common arm analysisshows great similarities in patient demographics betweenFACS and S0003. (2) Variable toxicities maybe due to differences in cumulative paclitaxel dose(neuropathy) and/or population-related pharmacogenomics(increased neutropenia and febrile neutropeniain FACS despite lower paclitaxel dose).(3) Survival is increased in FACS. (4) Future jointJapan-US clinical trials should consider possiblepharmacogenomic differences in drug disposition.Abstract #7041Influence of Gender on TreatmentOutcome and Toxicity in Small-CellLung CancerS. Singh, W. Parulekar, N. Murray, R. Feld, B. Evans, D. Tu,J. Pater, F. A. ShepherdPrincess Margaret Hospital, Toronto, Ontario, Canada; NationalCancer Insitute of Canada, Kingston, Ontario, Canada; BritishColumbia Cancer Agency, Vancouver, British Columbia,Canada; Cancer Care Ontario, Toronto, Ontario, CanadaFemale gender has been shown consistently to bea favorable prognostic factor in small-cell lung cancer(SCLC). Studies have shown that women withother tumor types experience greater treatment toxicity,but there have been few studies of gender-relatedtoxicity in SCLC.This was a gender-based retrospective analysis offour SCLC trials that were conducted by the NCICCTG between 1981 and 1996. All 1,006 patients (648male, 358 female) received similar chemotherapyconsisting of cyclophosphamide/doxorubicin/vincristineand etoposide/cisplatin. Toxicities examinedincluded myelosuppression, stomatitis, vomiting, andinfection. Other end points included number of dosereductions required, number of omitted cycles, responserates, and overall survival. Toxicities betweenthe genders were compared using the chi-square testin univariate analyses and logistic regression adjustingfor age, BSA, performance status, lactate dehydrogenase,and individual trial in multivariateanalyses.Women experienced significantly more toxicity inboth univariate and multivariate analyses (see table).However, toxic death rates were similar for men andwomen (1.5% vs 1.1%, P = .58). Despite increasedtoxicity, 76% of females vs 73.4% of males receivedall six treatment cycles (P = .38), but 52% of femalesvs 43.4% of males had treatment delayed for ≥ 2weeks (P = .022). The ORR was 80.3% for femalesand 66.9% for males (P < .0001) and the mediansurvival was 1.31 years for females and 0.91 formales (P < .0001).CONCLUSION: Women clearly experience morechemotherapy-related toxicity in the treatment ofSCLC, but this does not result in more toxic deaths oromitted treatment cycles, nor does it compromiseoutcome. See Table 1 below.
Abstract #7216A Phase II Study of Pegfilgrastim toSupport ACE 14 Chemotherapyfor the Treatment of SubjectsWith Small-Cell Lung CancerR. Pirker, E. Ulsperger, K. Aigner, J. MessnerVienna Medical University, Vienna, Austria; KrankenhausLainz, Vienna, Austria; Krankenhaus der Elisabethinen, Linz,Austria; Landesklinik fr Lungenkrankheiten, Salzburg, AustriaThatcher et al (J Clin Oncol 18:395-404, 2000)showed that dose-dense ACE (doxorubicin [Adriamycin],cyclophosphamide, etoposide) with G-CSFallows delivery of "chemotherapy planned dose ontime" (CPDOT), leading to improved survival insubjects with extensive small-cell lung cancer (SCLC).The aim of this study was to show that pegfilgrastimcan also support ACE 14 CPDOT in this setting.All 30 subjects received ACE (doxorubicin 40mg/m2 and cyclophosphamide 1,000 mg/m2 IV day1, etoposide 120 mg/m2 IV day 1 and 240 mg/m2 podays 2 and 3) every 14 days for up to 6 cycles, witha single dose per cycle of pegfilgrastim 6 mg SC onday 4. A cycle was considered "on time" if it startedno more than 17 days after the start of the previouscycle, and "at planned dose" if 75% of drug wasadministered for each agent. For any given cycle,subjects were defined as having CPDOT if bothcriteria were satisfied.Twenty-seven subjects received pegfilgrastim andat least 1 cycle of ACE and 17 received all 6 cycles ofACE. Twenty subjects (74%) had no bone marrowinvolvement at baseline. The mean baseline ANCwas 7.4 * 109/L (range: 3.0-23.3); over the study,mean ANC levels at the start of each cycle were 1.5* 109/L and were comparable to baseline levels. All22 subjects who started cycle 2 received full CPDOT.Of the 121 cycles delivered over the study, 107 (88%)were full CPDOT. In addition, 18 subjects (67%)received all their cycles according to full CPDOT.Of the 22 subjects who were assessed for diseaseresponse, 2 (9%) had complete response, 15 (68%)partial response (with an overall response rate of77%), and 2 (9%) stable disease. Safety data wereconsistent with the underlying patient group. Nine(33%) and 6 (22%) subjects experienced hematologicand nonhematologic events of toxicity grade 3/4:anemia 15%, leukopenia 7%, thrombocytopenia 11%,and febrile neutropenia 15%. All other adverse eventswere single episodes.CONCLUSION: These results indicate that pegfilgrastimenables delivery of dose-intensified ACEchemotherapy every 14 days in SCLC patients.Abstract #7138Prospective Multicenter Phase II Trial ofDocetaxel and Vinorelbine WithFilgrastim Support in Subjects WithAdvanced Non-Small-Cell Lung CancerR. D. Page, F. P. Smith, G. F. Geils, C. L. Beall, M. B. Ross,M. Fridman, B. J. AllenTexas Cancer Care, Weatherford, Texas; Sibley MemorialHospital, Washington, DC; Charleston Hematology/Oncology,Charleston, South Carolina; Morgantown Internal Medicine/Oncology, Morgantown, West Virginia; AMF Consulting,Los Angeles, California; Amgen, Inc, Thousand Oaks,CaliforniaAlthough platinum-based chemotherapy (CT) regimensare the standard of therapy in first-line non-small-cell lung cancer (NSCLC), a subset of patientsare ineligible to receive such agents. Docetaxel andvinorelbine are each active agents in NSCLC, and whencombined may provide enhanced activity. The purposeof this study was to evaluate the safety and efficacy ofthis combination therapy when administered on a every-2-week schedule with filgrastim support.This multicenter, community-based study wasconducted in 10 sites that enrolled 61 CT-naivepatients with stage IIIB/IV NSCLC. Treatmentconsisted of vinorelbine 45 mg/m2 and docetaxel60 mg/m2 day 1 and filgrastim 5 μg/kg days 2-14repeated every 2 weeks for 8 cycles. Patients wereevaluated for response at the end of cycles 2, 5, and 8.Confirmed response was defined as complete orpartial response based upon two or more evaluations,while best response was defined as the best of thethree evaluations. We report response rate and Kaplan-Meier estimate (95%CI) of time to disease progressionand 1-year survival.Of 61 enrolled patients 42% (95% confidenceinterval [CI] = 30%-54%) achieved either completeresponse or partial response as best response;13% (95% CI = 9%-17%) of patients had a confirmedresponse. Median time to disease progressionwas 160 days. With a median follow-up timeof 14.2 months, the estimated median survivaltime is 14.1 months (95% CI lower bound, 8.5 mo),and the 1-year survival rate is 57% (95%CI = 44%-70%). Five patients (8%) experiencedfebrile neutropenia (ANC < 500/mm3 plus bodytemperature ≥ 38.2oC). Overall calculated doseintensity (delivered/planned dose * 100) was 95%for vinorelbine and 93% for docetaxel.CONCLUSION: The combination of docetaxeland vinorelbine can be delivered with high doseintensity with filgrastim support. Additionally in thismulticenter noncontrolled study completed in a communitysetting, an interesting 1-year and overallmedian survival is observed and supports furtherevaluation.Commentary on Abstracts #7223,
#7007, #7041, #7216, and #7138
Jeffrey Crawford, MDExtending the observations from the UHC databasepreviously discussed, Daniel and colleagues(ASCO abstract #7223) reviewed the risk and mortalityassociated with febrile neutropenia in lungcancer patients. What was striking from their resultswas that the overall inpatient mortality wasquite distinct from other solid tumor patients andapproached that of patients with hematologic malignancies,with an overall inpatient mortality of12.1%. Factors associated with risk of inpatientmortality included metastatic disease, hypotension,pneumonia, and gram-negative bacteremia. Increasingage, advanced disease, nutritional and functionalimpairments, and comorbidities, along with theseinfectious complications, led to both risk of prolongedhospitalization and increased risk of mortality.Based on these data, prospective studies arewarranted to evaluate strategies to reduce the riskof febrile neutropenia in the high-risk setting of lungcancer.To further refine patients at risk, Gandara et al(ASCO abstract 7007) compared a carbo-platin/paclitaxel regimen from the results of a Japanesestudy (FACS) and SWOG S0003. Both studiesused a carboplatin dose of AUC 6, with a paclitaxeldose of 200 mg/m2 in FACS and 225 mg/m2in S0003. Despite the slightly lower paclitaxeldose, the incidence of neutropenia occurred in69% of the Japanese population vs 26% of theNorth American population. The febrile neutropeniarate was 16% in the Japanese populationand 3% in the North American population. Thisvariable toxicity suggests the potential for population-related pharmacogenomic differences andrequires further study.In a retrospective review of four small-celllung cancer trials from the NCIC group, Singhand colleagues (ASCO abstract #7041) demonstratedthat women clearly experienced morechemotherapy-related toxicity in the treatment ofsmall-cell lung cancer than men. Women weresignificantly more likely to develop grade 3 or 4leukopenia and also experienced more anemia,stomatitis, and vomiting. In addition to increasedtoxicity, the overall response rate and mean survivalwas also greater for women compared withmen. These gender differences are striking, andagain provide clues to optimized chemotherapydosing while preemptively managing treatmenttoxicities in populations at higher risk.In this regard, two studies reported the potentialfor growth factor support to facilitate the useof dose dense chemotherapy. Pirker and colleagues(ASCO abstract #7216) reported the resultsof the phase II study of pegfilgrastim tosupport ACE 14-day chemotherapy for the treatmentof patients with small-cell lung cancer. Twothirds of the patients received all their cycles ofchemotherapy at full dose on time, demonstratingthe ability to deliver a dose-dense myelosuppressiveregimen every 2 weeks. The treatment outcomeswere encouraging.This strategy of dose-dense therapy may haveapplication for further study in earlier stages, notonly in small-cell lung cancer but perhaps innon-small-cell lung cancer as well. In that regard,Page and colleagues (ASCO abstract #7138)reported the results of a multicenter phase II trialof docetaxel and vinorelbine with filgrastim support.Again the results were encouraging in thesetting of advanced-stage non-small-cell lungcancer. Evaluations of these strategies in earlierstages of the disease may provide more definitiveresults of potential benefit.Neutropenia and Breast CancerAbstract #A-36Predicting Neutropenic Risk andReduced Chemotherapy Dose Intensityin Patients With Early-Stage BreastCancer: Results From a ProspectiveNationwide RegistryG. H. Lyman, J. Crawford, D. Wolff, E. Culakova,M. Poniewierski, D. C. Dale, for the ANC Study GroupUniversity of Rochester, Rochester, New York; DukeUniversity, Durham, North Carolina; University ofWashington, Seattle, WashingtonChemotherapy-induced neutropenia is frequentlyassociated with neutropenic complications (NC) resultingin reduced relative dose intensity (RDI), potentiallycompromising patient outcomes.Retrospective analyses have identified several riskfactors for NC and reduced RDI, but these studies arelimited by variable reporting and missing data. Toovercome these limitations, a prospective registry ofpatients with cancer at 137 centers has been established.More than 2,500 patients initiating a new chemotherapyregimen have been registered, including 617patients with early-stage breast cancer-525 of whomhave completed at least one cycle of chemotherapy.Primary outcomes include ANC < 1,000 (58%),neutropenic complications (ANC < 500 or infection)(42%), and RDI < 85%. Regimens include AC or ACT(53%), CMF (12%), and CAF (9%). Both unconditional(pretreatment) and conditional (based on firstcycle event) models for risk of NC and reduced RDIhave been developed.In unconditional models, significant predictors ofNC include anthracycline regimens (OR = 7.9) andprior chemotherapy (OR = 3.52) (model P < .001;R2 =0.15; C-statistic 0.62). In conditional models,significant predictors of NC include first cycle ANCnadir (OR = 8.98), first cycle infection (OR = 3.15),prior chemotherapy (OR = 10.02); and BSA > 2 m2(OR = 0.51) (model P< .001; R2 = 0.61; C-statistic0.86). In models for reduced RDI, stage of diseaseand practice center were also significant independentpredictors.CONCLUSION: Recent large practice pattern surveyshave shown that breast cancer patients oftenreceive reduced chemotherapy dose intensity, primarilydue to NC. A reliable tool to identify patients atrisk for NC should permit more rational use of targetedsupportive care and delivery of improved chemotherapydose intensity.Abstract #776Predicting the Risk of NeutropenicComplications and Reduced DoseIntensity in Patients With Early-StageBreast Cancer: Results From aProspective Nationwide RegistryG. H. Lyman, J. Crawford, D. C. Dale, D. A. Wolff,E. Culakova, for the ANC Study GroupUniversity of Rochester Medical Center, Rochester,New York; Duke University Medical Center, Durham,North Carolina; University of Washington Medical Center,Seattle, WashingtonChemotherapy-induced neutropenia is frequentlyassociated with neutropenic complications resultingin reduced relative dose intensity (RDI), potentiallycompromising patient outcomes. Retrospective analyseshave identified several risk factors but are limitedby variable reporting and missing data. Aprospective registry of chemotherapy patients at 137centers has been established. Both unconditional(pretreatment) models and models conditional onfirst-cycle events for risk of neutropenic complications and reduced RDI have been developed.Overall, 2,222 patients have been registered priorto chemotherapy, including 617 patients with earlystagebreast cancer (ESBC), of which 525 have completedat least one cycle of chemotherapy. Adverseoutcomes include ANC < 1,000 (58%), neutropeniccomplications (ANC < 500 or infection) (42%), andRDI < 85% (11.3%). Regimens include AC or AC-T(53%), CMF (12%), and CAF (9%).In unconditional models, significant predictors ofneutropenic complications include anthracycline regimens(OR = 7.9) and prior chemotherapy (OR =3.52) (model P < .001; R2 = 0.15; c-statistic 0.62). Inmodels conditional on first-cycle events, significantpredictors of neutropenic complications include firstcyclenadir (OR = 8.98), first-cycle infection(OR=3.15), prior chemotherapy (OR = 10.02), andbody surface area > 2 m2 (OR = 0.51) (model P < .001;R2 = 0.61; C-statistic 0.86). In models for reducedRDI, stage and practice center were also significantindependent predictors. Separate validation of thesemodels will be presented.CONCLUSION: Recent large practice surveys demonstratethat breast cancer patients often receive reducedchemotherapy dose intensity, primarily due toneutropenic complications. A reliable tool to identifypatients at risk for neutropenic complications shouldpermit more rational use of targeted supportive care anddelivery of optimal chemotherapy dose intensity.Abstract #677The Role of Growth Factor SupportFollowing Neutropenic Events inEarly-Stage Breast Cancer PatientsTreated With Adjuvant Docetaxel,Doxorubicin, and Cyclophosphamide:A Subanalysis of BCIRG 001C. L. Vogel, J. R. Mackey, M. Martin, on behalf of theBCIRG 001 InvestigatorsCancer Research Network, Plantation, Florida; Cross CancerInstitute, Edmonton, Alberta, Canada; Hospital UniversitarioSan Carlos, Madrid, SpainDocetaxel, doxorubicin, and cyclophosphamide(TAC) significantly improves disease-free and overallsurvival over FAC (hazard ratio 0.70 and 0.68,respectively), and is emerging as one of the mostactive adjuvant treatments in patients with node positiveearly-stage breast cancer (Martin M: SABCS2003 abstract #43). TAC is generally well tolerated,but is associated with a higher incidence of febrileneutropenia (FN) vs FAC. ASCO Clinical PracticeGuidelines recommend secondary prophylaxis withG-CSF after FN in a prior cycle to maintain doseintensity.Patients with node-positive breast cancer wererandomized to TAC (75/50/500 mg/m2 q3wk * 6) orFAC (500/50/500 mg/m2 q3wk * 6). Corticosteroidpremedication and prophylactic ciprofloxacin weregiven with TAC, but not with FAC. In case of FN(grade 2 fever with grade 4 neutropenia), patientswere treated with G-CSF for all subsequent cycles.This retrospective subgroup analysis compares theincidence of FN for cycles treated without and withG-CSF.A total of 1,491 patients were accrued, with 1,480evaluable for safety (TAC 744, FAC 736). A similarnumber of cycles were delivered in both arms (TAC4010, FAC 4007). Febrile neutropenia occurred asfollows: TAC, 183 patients (24.7% patients, 5.4%cycles); FAC 18 patients (2.5% pts, 0.5% cycles),with at least half the FN occurring in the first cycle(TAC: 97/183 patients, FAC 9/18 patients). G-CSFwas administered to 250 TAC patients and 93 FACpatients. Among these, G-CSF was used as secondaryprophylaxis for 87% (TAC) and 44% (FAC) of patients.The rate of FN (per cycle) without vs withG-CSF among all patients was as follows: TAC 187/3,114 (6.0%) vs TAC+G-CSF 28/896 (3.1%); FAC19/3,704 (0.5%) vs FAC+G-CSF 1/303 (0.3%). Therewere no septic deaths during treatment with eitherTAC or FAC, regardless of the use of GCSF.CONCLUSION: Among patients treated withTAC, the use of G-CSF decreased the incidence ofneutropenic complications, although it remained higherthan for pts treated with FAC. For patients experiencinga neutropenic event with TAC, secondaryprophylaxis with G-CSF is appropriate. The impactof G-CSF on other clinical safety parameters will alsobe presented.Abstract #619Dose-Limiting Effects of NeutropenicEvents in Six European Audits ofAdjuvant Breast Cancer ChemotherapyR. Leonard, T. D. Szucs, R. Pettengell, R. Paridaens,C. Jackisch, M. Constenla, A. Bosly, M. Schwenkglenks, forthe Impact of Neutropenia in Chemotherapy European StudyGroup (INC-EU)South West Wales Cancer Institute, Swansea, United Kingdom;European Center of Pharmaceutical Medicine, Basel,Switzerland; St George's Hospital, London, United Kingdom;University Hospital Gasthuisberg, Leuven, Belgium; UniversityHospital Marburg, Marburg, Germany; Complexo Hospitalariode Pontevedra, Pontevedra, Spain; CliniquesUniversitaires UCL, Godinne, BelgiumRetrospective audits of adjuvant breast cancerchemotherapy (CT) were performed in several Europeancountries. Results of a combined analysis of sixaudits from Austria, Belgium, Germany, Spain, andthe United Kingdom are reported.Variables available in all six datasets were mergedinto a dataset of individual observations and theirdefinitions were harmonized. We assessed the incidenceof neutropenic events and of low averagerelative CT dose intensity (ARDI). Adjusted oddsratios (ORs) of low ARDI occurrence were calculatedby robust multiple logistic regression. Neutropenicevents were defined as neutropenia-relatedhospitalization, dose reduction ≥ 15%, and/or dosedelay ≥ 7 days. Low ARDI was defined as ARDI ≤85%.A total of 2,633 patients had a mean age at diagnosis± standard deviation of 51.0 ± 11.3 years (interauditrange [IAR]: 48.0 ± 10.9 to 52.5 ± 11.7 years).Patients were postmenopausal in 51% of cases and64% were hormone receptor positive. The diagnosticspread was stage I 19%, II 65%, and III 16%. Fiftyeightpercent received CMF-based regimens, 39%anthracycline-containing, and 3% other regimens.Concomitant radiotherapy was reported in 32%, anduse of colony-stimulating factors (CSF) in 12%. In3%, CSF use started in cycle 1.Neutropenic events were observed in 20% of patients(IAR: 14%-27%). Repeated neutropenic eventswere seen in 8% (IAR 6%-11%). Low ARDI wasobserved in 14%. In those without and with neutropenicevents, low ARDI was observed in 6% vs 17%(P < .005). Low ARDI was independently associatedwith neutropenic event occurrence (OR 4.9, 95% CI3.6-6.7); use of a nonanthracycline regimen (OR 1.5,95% CI = 1.1-1.9); concomitant radiotherapy (OR1.3, 95% CI = 1.1-1.5); and disease stage. Odds ratiosfor stages II and III, compared to stage I, were 1.5(95% CI = 1.2-2.0) and 1.6 (95% CI = 1.2-2.3). Amore restrictive neutropenic event definition basedon the cell count data available lead to an OR of 2.6(95% CI = 1.7-3.7) for low ARDI, with the othercoefficients stable.CONCLUSION: Neutropenic events occurred in arelevant proportion of patients receiving breast cancerCT and showed a robust association with lowARDI, which may affect treatment outcomes. Ongoingprospective research should address the developmentof risk models to target preventive measuresand optimize CT.Abstract #583Reducing Dose Density in AdjuvantChemotherapy Is Detrimental inEarly Breast Cancer: A Review of872 Adjuvant Treatments in CentreFranois BaclesseT. Delozier, C. Sgura, C. Levy, C. Delcambre, D. Allouache,O. Switsers, J.-M. Ollivier, B. Vi, F. Joly, J.-Y. GnotCentre Franois Baclesse, Caen, FranceDose density is a key in the efficacy of chemotherapy.This concept has been tested in increasing dosedensity, especially in the adjuvant setting. Side effectsof chemotherapy such as leukopenia can inducedelay in dose administration, leading to a decrease indose density. The aim of the study was to evaluate theimpact of reducing dose density in adjuvant chemotherapyin early breast cancer. We compared disease-freesurvival and overall survival according to dose densityexpressed in treatment duration for women treated withthe same chemotherapy in our institution.We reviewed data from 872 women who receivedsix cycles of chemotherapy using the FEC60 regimen(5-FU 600 mg/m2, epirubicin 60 mg/m2,cyclophosphamide600 mg/m2 on day 1 repeated every 3 weeks).All estrogen receptor positive (ER+) patients receivedadjuvant tamoxifen for 5 years. According tothe protocol, chemotherapy administration was postponedfor 1 week if patients presented leukopenia onday 1. No reduction in the applied dose was performed.Overall, 181 patients (group 1) underwent their sixcycles of chemotherapy without any delay. In 343patients (group 2), chemotherapy was delayed up tothree times: treatment duration between 110 and 130days. Chemotherapy was delayed more than threetimes in 350 patients (group 3): treatment durationgreater than 130 days. The distribution of main characteristicswas similar in the groups. The medianduration of treatment was 125 days (range: 100-195days). Median age was 47 years (range: 25-69 years),66% were premenopausal, 63% node positive, 66%ER positive. The median follow-up is 4 years.Overall, 35, 75, and 117 patients relapsed in thethree groups, respectively, leading to a 4-year disease-free survival (DFS) rate of 79.4%, 76.0%, and66.1% respectively (P = 0.09). Patients who underwentchemotherapy in more than 130 days had ahigher risk of relapse (P = 0.03). No difference wasobserved in overall survival.CONCLUSION: We conclude that although nosurvival disadvantage is noted, the not respectingtime schedules in chemotherapy is detrimental inearly breast cancer patients.Abstract #552Impact of Reduced Dose Intensity ofAdjuvant Anthracycline-BasedChemotherapy in a Population-BasedCohort of Stage I/II Breast CancersA. Tinker, C. Speers, J. Barnett, I. Olivotto, S. ChiaBritish Columbia Cancer Agency, Vancouver, British Columbia,CanadaReductions in the dose intensity of adjuvant doxorubicinand cyclophosphamide (AC) chemotherapyin the treatment of early-stage breast cancer arefrequently required, with the impact on clinical outcomeuncertain. We examined whether a reduceddose intensity had an impact on relapse-free survival(RFS), breast cancer-specific survival (BCSS), oroverall survival (OS) in a population-based cohort ofearly-stage breast cancers treated with adjuvant AC.Women with stage I/II breast cancer treated withadjuvant AC (A: 60 mg/m2, C: 600 mg/m2 on a 21-dayschedule) between 1990 and 1995 were retrospectivelyidentified through the British Columbia CancerAgency (BCCA) pharmacy database and linked tothe BCCA Breast Cancer Outcomes Unit database. Adose reduction was defined as a reduction of at leastone of the chemotherapy agents by at least 25% in anygiven cycle. Dose delay was defined as a delay indelivering treatment by at least 5 days. Cases wereclassified into the following four cohorts: cohort 1,entire course of treatment delivered at full doses andon time; cohort 2, one single dose reduction or dosedelay; cohort 3, more than one dose reduction or dosedelay; cohort 4 = two cycles of chemotherapy delivered.No growth factor support was utilized in anycases.A total of 484 cases were retrospectively identified(cohort 1, n = 268; cohort 2, n = 88; cohort 3, n = 89;cohort 4, n = 39) with a median follow-up of 9.6 years.The four cohorts were well matched for most baselineprognostic factors except for slight imbalances inlymph node status (P = .05) and adjuvant hormonaltherapy (P = .05). Fifty-five percent of the entirecohort had node-positive disease. Overall 45% ofcases had a reduced dose intensity delivered. However,there were no significant differences in 8 year RFS(P = .94), BCSS (P = .87), and OS (P = .86) betweenthe four cohorts. The 8-year outcomes for cohorts 1through 4, respectively, were RFS (72%, 74%, 74%,68%), BCSS (80%, 77%, 82%, 80%), and OS (78%,76%, 80%, 77%).CONCLUSION: Although reductions in the doseintensity of adjuvant AC chemotherapy for earlystagebreast cancer was common, it did not appear tosignificantly impact on clinical outcomes in thiscohort of patients with stage I/II breast cancer.Abstract #589A Phase III Randomized TrialComparing the Tolerability ofDose-Dense Chemotherapy in Olderto That in Younger Breast CancerPatients With Four or More PositiveLymph Nodes
S. Kuemmel, J. Krocker, A. Kohls, M. Budner,G.-P. Breitbach, K. Ruhland, U. Khler, B. Reischauer,J.-U. Blohmer, D. EllingDepartment of Senology, Campus Charit Mitte, Berlin,Germany; Krankenhaus Lichtenberg, Berlin, Germany; Stdt.Krankenhaus Ludwigsfelde, Germany; Humaine-Klinikum,Bad Saarow, Germany; Stdt. Krankenhaus, Neunkirchen,Germany; Stdt. Klinikum, Stralsund, Germany; Stdt. Klinikum,Leipzig, GermanyRecently published data showed older patients tohave similar dose-related benefits in reducing breastcancer-related relapse and mortality to those ofyounger women. This analysis was performed toassess the feasibility of a prospective randomizeddose-dense regimen in the treatment of women overthe age of 60.From June 1996 to November 2000, 211 primarybreast cancer patients with four or more positivelymph nodes were prospectively randomized to eitherfour cycles of epirubicin and paclitaxel (ET) at 2-week intervals followed by three cycles of cyclophosphamide,methotrexate, and fluorouracil (CMF)with G-CSF support (group A) (age < 60 years [A1],n = 79; age ≥ 60 years [A2], n = 25), or four cycles ofepirubicin and cyclophosphamide (EC) at 3-weekintervals followed by three cycles of CMF (group B)(age < 60 years [B1], n = 80; age ≥ 60 years [B2], n= 27).The median age was 52 years (range: 26-59 years)in group A1, 64 years (range: 60-71 years) in groupA2, 48.5 years (range: 32-59 years) in group B1 and64 years (range: 60-72 years) in group B2. All sevencycles were administered in 95% (A1), 100% (A2),99% (B1), and 89% (B2) of patients. A dose reductionwas made only in 1% (A1), 2% (A2), 1% (B1), and1% (B2) of cycles. In comparing these four groups,delays in administering cycles were more frequent inthe dose-intensified group of older patients (6% in A1vs 17% in A2; 6% in B1 vs 11% in B2). The hematologictoxicity data are shown in Table 1. NCI-CTCgrade 3/4 nonhematologic toxicities were rare andrevealed no differences between the groups.
CONCLUSION: The administration of a dosedenseregimen with epirubicin and paclitaxel followedby CMF with G-CSF support is feasible inelderly patients with a tolerable safety profile. Referringto the analyzed data, a decreased hematopoieticpotency must be considered in older patients.
Commentary on Abstracts #A-36,
#776, #677, #619, #583, #552,and #589
Jeffrey Crawford, MD
Lyman and colleagues (MASCC abstract#A-36, ASCO abstract #776) reported anothervery important outcome of neutropenia distinctfrom infection-the potential for reduced chemotherapydose intensity in patients with early-stagebreast cancer. Utilizing the prospective ANCregistry that was described earlier, 600 patientswere evaluated with early-stage breast cancer. Inthis cohort, 42% of patients developed ANC lessthan 500 or infection. Predictors for neutropeniccomplications included anthracycline-based regimens,and prior chemotherapy. In models forreduced relative dose intensity, both stage ofdisease and practice center were also significantindependent predictors. By developing a tool toidentify patients at risk for neutropenic complications,the investigators hope to develop a morerational use of growth factor support to improvechemotherapy dose delivery.Until better risk models are developed forprimary prophylaxis, both colony-stimulatingfactors and secondary prophylaxis after a priorneutropenic event will remain a commonly usedstrategy to maintain dose intensity in early-stagebreast cancer. Vogel and colleagues (ASCO abstract#677) identified 250 patients who receivedG-CSF in support of TAC chemotherapy. Eightysevenpercent of those patients received G-CSFas a secondary prophylaxis. The rate of febrileneutropenia per cycle was 3.1% for patients receivingG-CSF with TAC compared to 6% ofpatients receiving TAC without the G-CSF. Theauthors concluded that for patients experiencingneutropenic events with TAC, secondary prophylaxiswith G-CSF is appropriate. (As noted inthe earlier section redefining the risk threshold offebrile neutropenia, TAC chemotherapy would bein the range of 20% and would qualify for primaryprophylaxis.)In another report from six European audits ofadjuvant breast cancer chemotherapy, Leonardand colleagues (ASCO abstract #619) found thatneutropenic events were clearly associated with alikelihood of low relative chemotherapy doseintensity. In this review of 2,633 patients, anaverage relative dose intensity of less than 85%was observed in 14% of the population. Six percentof patients without neutropenic events experienceda low relative dose intensity vs 17% ofpatients who did experience a neutropenic event
(P < .
005).And what is the evidence that reducing doseintensity in adjuvant chemotherapy is detrimentalto outcome in early-stage breast cancer? Toaddress this, Delozier and colleagues (ASCOabstract #583) identified three groups of patientsundergoing adjuvant FEC chemotherapy. Accordingto protocol, chemotherapy was postponedfor 1 week in patients who presented withleukopenia on day 1. In their study of 872 patients,181 patients underwent all six cycles of therapywithout delay (group 1). A total of 343 patients ingroup 2 experienced up to three treatment delays.Group 3 experienced more than three treatmentdelays and this accounted for 350 patients. The4-year disease-free survival rate was 79.4%, 76%,and 66.1%, respectively, for the three groups.While the differences did not reach statisticalsignificance (
P =
.09), the trend suggests thatdelivery of chemotherapy on time is an importantvariable. Furthermore, patients whose total treatmenttime exceeded 130 days did have a higherrisk of relapse (
P =
.03).This provocative data, along with previousstudies by Bonadonna and the Cancer and LeukemiaGroup B, continues to emphasize the importanceof chemotherapy dose delivery as animportant predictor of outcome. Prospective trialswith CSF support to maintain full-dose chemotherapyon time would provide definitivesupport of this principle.The importance of such a prospective trial isevidenced by other data from Tinker and colleagues(ASCO abstract #552). This group wasunable to demonstrate that reductions in the doseof adjuvant AC chemotherapy in early breastcancer were associated with an impact on clinicaloutcomes. This retrospective analysis involved484 patients receiving AC chemotherapy regimen.Whether or not the differences between thisCanadian trial and the French study are due todifferences in population, sample size, or thechemotherapy regimen can only be further definedby prospective studies.Lastly a phase III randomized trial was performedto compare the tolerability of dose-densechemotherapy in older compared to youngerbreast cancer patients (ASCO abstract #589).Kuemmel and colleagues from Germany wereable to demonstrate that four cycles of epirubicinand paclitaxel could be administered every 2weeks along with three cycles of CMF chemotherapywith G-CSF support. Across a variety ofsubgroups of patients, only 1% to 2% of patientsrequire dose reduction. In the older population,dose delays occurred more commonly duringdose-dense therapy: 17% vs 6%. There were alsomore frequent dose delays in the older patientsreceiving every-3-week chemotherapy: 11% vs6%. However, overall the authors concluded thata dose-dense chemotherapy regimen with G-CSFsupport is feasible in the elderly with a tolerablesafety profile. Thus, in settings where dose-densestudies have been shown to be more efficacious,older women with breast cancer can be consideredfor such a strategy when appropriate.
Neutropenia and Lymphoma
Abstract #8068
A Model to Predict Chemotherapy-Related Severe or Febrile Neutropeniain Cycle 1 Among Breast Cancer andLymphoma Patients
V. A. Morrison, V. Caggiano, M. Fridman, D. J. DelgadoVA Medical Center, Minneapolis, Minnesota; Sutter CancerCenter, Sacramento, California; AMF Consulting, Los Angeles,Calfornia; Amgen Inc, Thousand Oaks, California
Chemotherapy used to treat cancer may producesevere (absolute neutrophil count ≤ 250/mm
3
) orfebrile neutropenia, which often results in fever,infection, and hospitalization. This can lead to dosedelays or reductions in subsequent chemotherapycycles and/or early termination of therapy. Recentstudies suggest most patients who experience severefebrile neutropenia do so early in the course ofchemotherapy, in particular during the first cycle.Several recent risk models for neutropenia haveidentified baseline patient characteristics that predictthe occurrence of neutropenia. The ability to identifypatients at risk for developing neutropenia early intheir therapy might help guide appropriate hematopoieticgrowth factor use. We evaluated possible riskfactors associated with cycle 1 severe febrile neutropeniaamong a sample of patients with non-Hodgkin'slymphoma (NHL) or breast cancer.A historical case series of 1,617 patients (704 NHLand 913 early-stage breast cancer) who receivedinitial chemotherapy at 16 community and academiconcology practices between 1991 and 1999 wereselected for study. Severe febrile neutropenia wasdefined as an absolute neutrophil count ≤ 250/mm
3
orfebrile neutropenia.A total of 461 patients (29%) experienced at leastone episode of severe febrile neutropenia; 268 (58%)of these patients (167 [59%] with NHL and 101[56%] with breast cancer) had severe febrile neutropeniain cycle 1. Risk factors associated with cycle 1severe febrile neutropenia included age ≥ 65 years(odds ratio [OR] 2.08; 95% CI = 1.48-2.92); baselinehemoglobin < 12.0 g/dL (OR 1.90; 95% CI = 1.41-2.58); presence of heart, renal, or liver disease (OR2.12; 95% CI = 1.03-4.36); NHL (OR 1.64; 95% CI= 1.16-2.32); planned full chemotherapy dose intensity(OR 2.74; 95% CI = 1.55-4.84); and no growthfactor in the first 5 days of cycle 1 (OR 1.82; 95% CI= 1.07-3.08).CONCLUSION: Data routinely available to theclinician can help identify patients at risk for severefebrile neutropenia in cycle 1. In our model assessingchemotherapy-related severe febrile neutropenia inbreast cancer and lymphoma, patients ≥ 65 weretwice as likely to have severe febrile neutropenia incycle 1.
Abstract #A-60
Prevalence and Predictors of FebrileNeutropenia in Patients WithAggressive Non-Hodgkin'sLymphoma
D. C. Dale, G. H. Lyman, J. Crawford, for the ANC StudyGroupUniversity of Washington, Seattle, Washington; University ofRochester, Rochester, New York; Duke University, Durham,North Carolina
Febrile neutropenia (FN) is a serious complicationof chemotherapy that frequently requires hospitalization,dose reductions, and dose delays. As a consequence,FN may potentially compromise overall doseintensity and thus negatively impact long-term survivaland the potential for cure.To evaluate the delivery of chemotherapy doseintensity, the incidence of FN, and the impact ofdemographic and clinical characteristics, we conducteda nationwide survey of 567 community oncologypractices and analyzed the records of 4,522patients with aggressive non-Hodgkin's lymphoma(NHL) receiving standard regimens of CHOP, CHOPrituximab(Rituxan), or CNOP. The average age ofthe study population was 61 years, with 58% ofpatients > 60 years of age at diagnosis. Overall, 53%of patients (n = 2,401) received relative dose intensity< 85% and 21% (n = 945) experienced at least oneepisode of FN, 76% (n = 717) of whom requiredhospitalization. Febrile neutropenia was more prevalentin patients > 60 years old, BSA > 2, ECOGperformance status ≥ 2, stage ≥ 3, International PrognosticIndex ≥ 2, and those receiving CHOP-rituximab.In multivariate analysis, significant predictors ofFN included age > 60 (OR = 1.57), female gender (OR= 1.32), stage ≥ 3 (OR = 1.30), ECOG performancestatus ≥ 2 (OR = 1.88), and baseline neutrophil counts< 1,000 (OR = 1.53). Lower risk of FN was observedover time and in those receiving primary CSF prophylaxis(OR = 0.84). Older age was a significant riskfactor for FN when no prophylactic CSF was used(OR = 1.80), but not when prophylactic CSF was used(OR = 0.77).CONCLUSION: These data illustrate a high prevalenceof FN of about 20% in patients with NHL,most of whom were treated with reduced chemotherapydose intensity.
Abstract #6599
Febrile Neutropenia and Reduced DoseIntensity in Patients With AggressiveNon-Hodgkin's Lymphoma TreatedWith CHOP and CNOP
D. C. Dale, J. Crawford, O. Agboola, G. H. Lyman, forthe ANC Study GroupUniversity of Washington, Seattle, Washington; DukeUniversity Medical Center, Durham, North Carolina;University of Rochester Medical Center, Rochester, New York
Myelosuppression remains the major doselimitingtoxicity of systemic chemotherapy in patientswith intermediate-grade non-Hodgkin's lymphoma(NHL). Recent studies support the importanceof sustaining dose intensity in this setting.A survey of 1,243 community oncology practices,including nearly 5,500 patients receiving chemotherapyfor NHL, was undertaken to evaluate the impactof demographic, clinical-, and treatment-related factorson delivered dose intensity. Relative dose intensity(RDI) was estimated for each drug as the ratio ofdose intensity received to standard dose intensity foreach regimen.This analysis is limited to 3,536 patients withaggressive NHL (Working Formulation D-H) treatedwith CHOP (87%) or CNOP (13%) chemotherapyover a 10-year period from 1992 to 2001. Averageage was 61 years, with 45% > 65 years and 54% male.Diffuse large cell histology was reported in 57%.Febrile neutropenia occurred one or more times in21% of patients of which 76% required hospitalization.In multivariate analysis, significant risk factorsfor febrile neutropenia included age ≥ 65 (OR = 1.8),female gender (OR = 1.5), heart disease (OR = 3.3),liver disease (OR = 3.4), and ECOG performancestatus ≥ 2 (OR = 2.1).Average RDI to reference standard was 84%, withaveraged planned and unplanned reductions in RDIof 4.5% and 11.5%, respectively. Forty-five percentof patients received average RDI < 85%. Risk factorsfor average RDI ≤ 85% included age ≥ 65 (53%),stage 4 (53%), ECOG performance status ≥ 2 (57%),female gender (48%), body mass index (BMI) ≥ 30(50%), and hospitalization for febrile neutropenia(54%). Colony-stimulating factors were used in 41%of patients primarily in response to a previous episodeof febrile neutropenia. In multivariate analysis,significant independent risk factors for average RDI< 85% included age > 65 (OR = 1.8), female gender(OR = 1.2), stage 4 (OR = 1.6), CHOP (OR = 1.3),BMI ≥ 30 (1.4) and previous febrile neutropeniahospitalization (1.4).CONCLUSION: Nearly half of patients with aggressiveNHL histology treated with CHOP-like regimensexperienced substantial dose reductions relatedto age, stage, gender, obesity, regimen, and previousfebrile neutropenia hospitalization.
Commentary on Abstracts #8068,#A-60, and #6599
Jeffrey Crawford, MD
In a study that overlaps both breast cancer andlymphoma, Morrison and colleagues (ASCO abstract#8068) developed a model to predict chemotherapyrelated severe or febrile neutropeniain cycle 1 among both breast cancer and lymphomapatients. Among 461 patients, 29% experiencedat least one severe or febrile neutropenicepisode. Of these episodes, 58% occurred in thefirst cycle of treatment. The risk factors identifiedfor cycle 1 neutropenic complications includedage greater than 65, baseline hemoglobin lessthan 12 g/dL, the presence of comorbidities relatedto heart, renal or liver disease, planned fulldosechemotherapy dose intensity, and no growthfactor support in the first 5 days of cycle 1. Thesedata come from a historical case series from 16community and academic oncology practices andcan be validated in other ongoing databases.In the abstracts by Dale and colleagues(MASCC abstract #A-60, ASCO abstract #6599),the prevalence and predictors of febrile neutropeniain patients with aggressive non-Hodgkin'slymphoma were evaluated from a nationwidesurvey of 567 community oncology practices thatidentified 4,522 patients with aggressive non-Hodgkin's lymphoma receiving CHOP or aCHOP-like regimen. The prevalence of febrileneutropenia in this overall population was approximately20%, which would place CHOP as aregimen to be considered for a primary prophylaxisbased on redefined risk threshold previouslydiscussed. Furthermore, an even higher risk offebrile neutropenia occurred in patients over 60,those of female gender, those patients with stageIII or IV disease, and those patients with anEastern Cooperative Oncology Group performancestatus of 2 or greater, as well as patientswith a pretreatment absolute neutrophil count ofless than 1 * 109/L. Patients with a body surfacearea greater than 2 m2 and patients who startedprophylactic CSF in the first cycle were at areduced risk of febrile neutropenia.Thus, it appears that these two different studiesin lymphoma have resulted in several similarcharacteristics that increased the patient's riskfor febrile neutropenia. If a risk threshold of 20%is used, most patients with aggressive lymphomareceiving a CHOP-like regimen will qualify forfirst-cycle prophylaxis. It is clear that these risksare even further elevated in the patient populationspreviously described.
Other Uses of Colony-Stimulating Factors
Abstract #6639
Pegylated Filgrastim After High-DoseChemotherapy and Autologous StemCell Transplant
M. H. Jagasia, J. Greer, A. Kassim, S. Mineishi, D. Morgan,K. Ruffner, F. SchueningVanderbilt University Medical Center, Nashville, Tennessee
Pegylated filgrastim (PF) (6 mg fixed dose) isequivalent to daily filgrastim (5 μg/kg/d for 14 days)after chemotherapy in decreasing the duration ofneutropenia. Daily filgrastim (5 μg/kg) started afterhigh-dose chemotherapy (HDC) and autologous stemcelltransplant (ASCT) leads to significant decreasein time to neutrophil engraftment. The role of PF afterHDC and ASCT is not known. We propose to studythe role of PF given as a single fixed dose (6 mg)subcutaneously (SC) on day +1 after stem-cell infusion.A total of 15 patients undergoing HDC and ASCTusing peripheral blood stem cells for multiple myelomaor lymphoma have been enrolled in an ongoingstudy. Stem cells were collected with peripheralblood pheresis after cyclophosphamide (3 g/m
2
) anddaily filgrastim (10 μg/kg). All patients were eligible for HDC and ASCT as per institutional criteria, and consented. The conditioning regimens were highdosemelphalan (200 mg/m
2
) for myeloma, and cyclophosphamide(7,200 mg/m
2
), BCNU (400 mg/m
2
), and infusional etoposide (2,400 mg/m
2
) for lymphoma.Patients received anti-infective prophylaxiswith acyclovir, levofloxacin, and flucanozole startingday -1. Packed red blood cells (PRBC) andplatelets were administered for a hematocrit andplatelet count of less than 25% and 20,000/μL,respectively.There were no adverse events attributable to PF.All patients engrafted neutrophils and platelets. Themedian time to NE was 10 days (range: 8-11 days).Incidence of febrile neutropenia was 67%, with amedian duration of 2 days (range: 1-4 days). Twentythreebacterial blood cultures were obtained in 10patients for febrile episodes or a suspected line infection,with only 5 (in 5 patients) being positive, all forcoagulase-negative staphylococcus. Antibacterialsother than for prophylaxis was required in 60% (9/15)of patients. Antifungals other than for prophylaxiswere not required. The median time to platelet engraftmentwas 17 days (range: 14-19 days). Themedian number of transfusions was 3 units of PRBCand 12 units of platelets. The median days withmucositis and total parenteral nutrition were 3 (range:0-9 days) and 3 (range: 0-12 days), respectively.CONCLUSION: Pegylated filgrastim at a fixeddose of 6 mg SC administered on day +1 after HDCand ASCT appears to be equivalent to daily filgrastim.
Abstract #6634
GM-CSF + Interferon-Alpha Inducea Graft-vs-Leukemia Effect in BoneMarrow Transplant Patients WithRelapsed AML and ALL
M. Arellano, E. K. WallerEmory University, Atlanta, Georgia
Relapse of acute leukemia following allogeneichematopoietic progenitor cell transplantation (HPCT)has a poor prognosis. Second HPCT and donor leukocyteinfusions have significant toxicity with limitedeffectiveness. GM-CSF and interferon-alfa (IFN)activate dendritic cells and induce leukemia cells toexpress costimulatory molecules and enhance alloantigenpresentation, potentially inducing graft-vsleukemiaeffects. We hypothesized GM-CSF andIFN induce antileukemic effects in patients withrelapsed acute leukemia.We performed a single-institution retrospectivestudy of 97 patients with AML or ALL relapsed afterallogeneic HPCT. The dose and duration of cytokines,toxicities, response, and postrelapse survivalwere analyzed. Overall survival rates of cytokinetreatedpatients were compared with relapsed leukemiapatients who did not receive cytokineimmunotherapy.Six patients received GM-CSF and IFN injectionsfollowing the diagnosis of ALL or AML relapsedafter allogeneic HPCT. The median dose of GM-CSFwas 500 ?g administered 3 times per week (median ofsix doses). The median dose of IFN was 3 millionunits administered 1 to 3 times per week (median ofsix doses). Four of the six patients (67%) exhibited apathologic or hematologic remission with an averageresponse of 3.3 months (range: 6 weeks to 12 months)Toxicities included malaise, myalgias, and fever.Graft-vs-host disease was documented in four of thesix patients, and occurred at a median of 6 weeks afterbeginning cytokine immunotherapy.The median survival of cytokine-treated patientswas 130 days, with three of six patients remainingalive at a median follow-up of 373 days. Two of threesurviving patients remain without evidence of diseaseon no immunosuppressive drugs. Cytokinetreatedpatients had better survival compared to 91noncytokine-treated patients with relapsed ALL orAML (median survival of 52 days, 1-year survival of8%,
P
= .02).CONCLUSION: The administration of GM-CSFand IFN may induce remissions in patients with acuteleukemia who have relapsed after allogeneic HPCT.Further studies to evaluate the optimal dosing scheduleand duration of therapy should be pursued.
Abstract #6612
Priming GM-CSF and Low-DoseCytarabine in the Treatment ofHigh-Risk and Elderly AcuteMyelogenous Leukemia andMyelodysplastic Syndrome Patients
E. S. Winer, K. B. Miller, G. W. ChanTufts-New England Medical Center, Boston, Massachusetts;Beth Israel Deaconess Medical Center, Boston, Massachusetts
Priming of leukemic cells with cytokines mayenhance the efficacy of cell cycle chemotherapy. Weattempted to use GM-CSF to enhance the effects oflow-dose cytarabine to increase cell recruitment, andoptimize cell cycle chemotherapy in patients unableto tolerate conventional induction chemotherapy.We evaluated the efficacy of GM-CSF primingwith low-dose cytarabine and concomitant hydroxyureain high-risk, elderly AML and advanced MDSpatients. Patients received induction chemotherapywith GM-CSF 250 μg/m
2
/d by continuous infusiondays 1-7, hydroxyurea 500 mg po qid day 1 and500 mg po tid days 2-15, and cytarabine 20 mg/m
2
/dby continuous infusion days 2-15.Forty-nine patients (26 female, 23 male) weretreated, median age 68.5 years (range 48-85); 78%had no previous treatment, and 22% of patients had amedian of two prior treatments (range: 1-7). All patientswere not eligible for standard induction chemotherapy.Following treatment, 44% of the assessablepatients had documented aplastic bone marrows.The overall response rate was 49%; 33% with CRand 16% with PR. Median duration of CR was 251days (range: 36-842 days). The CR for secondaryAML was 18%, for de novo AML 38%, and for MDS67%. None of the patients developed mucositis oralopecia, and no patient had greater than grade Inausea and vomiting. The most common adverseeffects were neutropenic fever (68%), atrial fibrillation(7%), and congestive heart failure (5%). Fivepatients died of treatment-related toxicity (sepsis).The median survival for all patients was 150 days, forpatients achieving CR was 290 days, and the 1-yearsurvival rate was 24%.CONCLUSION: GM-CSF priming can enhancethe cytoreductive effects of low doses of cytarabine.This combination therapy is well tolerated, and shouldbe explored as an alternative regimen for high-riskand elderly AML and advanced MDS patients.
Commentary on Abstracts #6639,#6634, and #6612
Jeffrey Crawford, MD
This commentary has focused on our understandingof the role of colony-stimulating factors,particularly G-CSF and pegfilgrastim, in the managementof chemotherapy-induced neutropenia-both to reduce the risk of neutropenic complicationsand to improve chemotherapy dose deliveryin appropriate settings. For the readership, however,it should be noted that the impact of colonystimulatingfactors and potential applicationsextend to many other therapeutic settings.In ASCO abstract #6639, Jagasia and colleaguesdescribe the use of pegfilgrastim afterhigh-dose chemotherapy in autologous stem celltransplant. In this setting, a single fixed dose of 6mg SC administered on day + 1 after high-dosechemotherapy and autologous stem-cell transplantappeared to be equivalent to prolongeddaily filgrastim. Thus, even in a setting of prolongedneutropenia, the steady-state concentrationsof pegfilgrastim during neutropenia resultin this very favorable pharmacodynamic effect.In addition, GM-CSF was evaluated as part ofa therapeutic approach, along with interferonalfa, to induce a graft-vs-leukemia effect in bonemarrow transplant with relapsed AML and ALL.In this study by Arellano (ASCO abstract #6634),this therapeutic combination induced remissionsin patients with acute leukemia after relapse fromallogeneic transplant, and provides an encouragingstrategy.Another abstract looked at priming with GMCSFand low-dose cytarabine in the treatment ofelderly AML and myelodysplastic syndrome patients.This study by Weiner and colleagues (ASCOabstract #6612) demonstrated that GM-CSF primingcould enhance cytoreductive effects of lowdoses of cytarabine. These two small trials suggestsome interesting and novel directions for theuse of GM-CSF. As further studies evolve, theseapplications of GM-CSF may provide an importantniche for this multipotent cytokine.
The author(s) have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
1.
Crawford J: One year of supportive cancer therapy. SupportiveCancer Therapy 1(4):199-200, 2004.
2.
Sejl BH, Walsh TJ, Holland SN: Infections in the cancerpatient, in DeVita VT et al (eds): Cancer: Principles and Practiceof Oncology, 6th ed, pp 2815-2868. Philadelphia, Lippincott,2001.
3.
Yowell S, Dmitri GD, Crawford J: Hematopoietic growthfactors, in Pazdur R et al (eds): Cancer Management: A MultidisciplinaryApproach, 8th ed, pp 883-898. Manhasset, CMP HealthcareMedia, 2004.
4.
Holmes FA, O’Shaughnessy JA, Vukeljas et al: Blinded,randomized, multicenter study to evaluate single administrationpegfilgrastim once per cycle versus daily filgrastim as the adjunctto chemotherapy in patients with high risk stage II or stage III/IVbreast cancer. J Clin Oncol 20:727-731, 2002.
Oncology Peer Review On-The-Go: Cancer-Related Fatigue Outcome Measures in Integrative Oncology
September 20th 2022Authors Dori Beeler, PhD; Shelley Wang, MD, MPH; and Viraj A. Master, MD, PhD, spoke with CancerNetwork® about a review article on cancer-related fatigue published in the journal ONCOLOGY®.