the Departments of Medical Oncology, Epidemiology and Biostatistics, Pulmonology, and Medical Technology Assessment, Radboud University Nijmegen Medical Centre
    Rijnstate Hospital, Arnhem
    Jeroen Bosch Hospital, GZG 's-Hertogenbosch
    Hospital Gelderse Vallei Ede
    Hospital Koningin Beatrix Winterswijk
    Canisius Wilhelmina Hospital Nijmegen
    Diaconessenhuis Meppel
    Trial Office Comprehensive Cancer Centre E, Nijmegen, the Netherlands

    ABSTRACT

    PURPOSE: Febrile neutropenia (FN) is a major complication of chemotherapy. Antibiotics as well as granulocyte colony-stimulating factor (G-CSF) are effective in preventing FN. This multicenter randomized phase III trial determines whether the addition of G-CSF to antibiotic prophylaxis can further reduce the incidence of FN in patients with small-cell lung cancer (SCLC) at the risk of FN.

    PATIENTS AND METHODS: Patients (N = 175) were stratified for stage of disease, performance status, age, and prior chemotherapy treatment, and were randomly assigned for treatment with cyclophosphamide, doxorubicin, and etoposide (CDE), followed by prophylactic antibiotics alone (ciprofloxacin and roxithromycin) or by antibiotics in combination with G-CSF on days 4 to 13.

    RESULTS: In cycle 1, 20 patients (24%) in the antibiotics group developed FN compared with nine patients (10%) in the antibiotics plus G-CSF group (P = .01). In cycles 2 to 5, the incidences of FN were practically the same in both groups (17% v 11%). Only the treatment parameters (odds ratio, 0.33; 95% CI, 0.14 to 0.78) and age (1.067 per year; 95% CI, 1.013 to 1.0124) were related to the probability of FN in cycle 1.

    CONCLUSION: Primary G-CSF prophylaxis added to primary antibiotic prophylaxis is effective in reducing FN and infections in SCLC patients at the risk of FN with the first cycle of CDE chemotherapy. For patients with similar risk of FN, the combined use of prophylactic antibiotics plus G-CSF can be considered, specifically in the first cycle of chemotherapy.

    INTRODUCTION

    Febrile neutropenia (FN) is still a major threat to patients treated with chemotherapy, resulting in loss of quality of life and even death. Several risk factors have been identified to assess the risk of FN in an individual patient. These risk factors are related to the patient (age, performance status) and underlying disease (extent, comorbidity), but also to the chemotherapy regimen used.1-4 To prevent chemotherapy-related FN, prophylactic antibiotics and granulocyte colony-stimulating factor (G-CSF) have been applied successfully.

    During the last decade, fluoroquinolones have been used increasingly and a meta-analysis in 1,408 neutropenic patients showed a significant reduction in the incidence of Gram-negative bacterial infections (relative risk [RR], 0.21), total infections (0.54), and episodes with fever (0.85).5 Clinically documented infections and infection-related deaths were not reduced by quinolone prophylaxis. A combination of a quinolone with prophylaxis directed against Gram-positive bacilli (penicillin, vancomycin, or macrolides) compared to quinolones alone, significantly reduced bacteremia (RR, 0.65), streptococcal infections (RR, 0.45), and, to a smaller extent, the incidence of FN (RR, 0.98), but without affecting the incidence of fever-related morbidity or mortality.6 Prophylactic ciprofloxacin plus roxithromycin versus two placebos reduced the incidence of FN, the number of infections, the use of therapeutic antibiotics, and hospitalizations due to FN by approximately 50%, with a reduced number of deaths due to infection in patients with small-cell lung cancer (SCLC) treated with CDE-chemotherapy (cyclophosphamide, doxorubicin, and etoposide).2 The prophylactic use of ciprofloxacin and roxithromycin versus placebo also proved to be cost-effective, and in some cases even cost-sparing.7 On the other hand, there is concern regarding the emergence of resistant microorganisms when prophylactic antibiotics are administered.8

    G-CSF prophylaxis shortens the duration of chemotherapy-induced neutropenia, resulting in a decreased incidence of FN, hospitalization, and the use of intravenous therapeutic antibiotics by approximately 50%.1,9 In 2000, the American Society of Clinical Oncology (ASCO) Growth Factors Expert Panel recommended G-CSF use for primary prophylaxis to be reserved for patients considered at high risk for FN or for prolonged hospitalization. This advice is based on data failing to demonstrate an improvement in clinical outcomes (complications of FN), tumor response, or survival as well as economic savings if the expected incidence of FN is more than 40% (depending on local cost factors).10,11

    Direct comparisons of the efficacy of either prophylactic strategy are not available. In the United States, there was considerable opposition to direct evaluation of the clinical and economic impact of both of these treatments, which necessitated the Cancer and Leukemia Group B (CALGB) to withdraw its proposed CALGB 9111 clinical trial studying G-CSF versus ciprofloxacin as alternative strategies against FN for patients with lung cancer, despite this study's support from the National Cancer Institute.12 Adams et al13 provide a nice summary of the relevant issues related to G-CSF for SCLC. The article raises concern over the overall costs and benefits of G-CSF during standard-dose chemotherapy. In fact, they conclude that the findings of a decade of health services studies have shifted toward not being supportive of colony-stimulating factor use for primary or secondary prophylaxis for SCLC patients. However, they also stress that some patients with a substantial increased risk of FN might still benefit from prophylaxis and that research should focus on identifying these patients.

    As of August 2005, no data of randomized trials were available that indicated whether the addition of G-CSF to primary antibiotics prophylaxis can further reduce the incidence of FN in patients at risk. In this article, we report on the results of a randomized trial in which the prophylactic role of G-CSF, added to antibiotics (ciprofloxacin plus roxithromycin) for the prevention of FN, was evaluated in SCLC patients treated with CDE chemotherapy, who were at risk of FN. CDE is, especially in Europe, one of the standard chemotherapy regimens in the treatment of extensive disease (ED) SCLC,14,15 with an incidence of FN from 29% to 53%.1,2,9 Our goal was to conduct a pragmatic trial with broad entry criteria that did not require specialized staging procedures or follow-up, thus providing results widely applicable to patients with cancer and their healthcare providers. This report focuses on the clinical results; an economic evaluation was also part of the study but will be reported separately.

    PATIENTS AND METHODS

    Patient Selection

    Patients needed to meet the following criteria to be included in the study: age older than 18 years, histologically or cytologically proven SCLC, either ED with a Karnofsky Performance Score (KS) higher than 40% or limited disease (LD) in combination with at least one of the following characteristics, KS 40% to 70%, age older than 60 years, or had to be diagnosed unsuitable for combined chemoradiotherapy. Previous treatment with chemotherapy or radiotherapy if patient had recovered from acute toxicity, cerebral metastasis, pancytopenia caused by bone marrow involvement, and previous malignancy if in complete remission were allowed. Patients were excluded in case of active uncontrolled infection, inadequate renal or hepatic function, any evidence or history of hypersensitivity or other contraindications to the drugs investigated in this trial. The ethical committee of each institution approved the investigational protocol. Written informed consent was obtained from each participating patient.

    Treatment

    Patients were stratified for stage of disease, performance status, age, and previous treatment with chemotherapy since ED, KS lower than 80%, age  60 years, and second-line therapy were all considered risk factors of FN. Subsequently, patients were randomly assigned for treatment with CDE chemotherapy followed by prophylactic antibiotics alone or antibiotics in combination with G-CSF.

    CDE chemotherapy consisted of cyclophosphamide 1,000 mg/m2 on day 1, doxorubicin 45 mg/m2 on day 1 intravenously, and etoposide 100 mg/m2 intravenously on days 1 to 3 every 3 weeks for five cycles. Antibiotic prophylaxis consisted of ciprofloxacin 500 mg bid and roxithromycin 150 mg bid orally, from days 4 to 13. If the patient was randomly assigned to the experiment arm, G-CSF was given subcutaneously on days 4 to 13 at a dose of 300 μg/d to patients whose body weight  75 kg or 480 μg/kg/d for patients whose body weight was more than 75 kg.

    When FN occurred, no dose reduction was applied to additional cycles of chemotherapy, unless FN was complicated by septic shock or absolute neutrophil count (ANC) remained less than 0.5 x 109/L for 7 days or longer. In case of platelet nadir count of less than 25 x 109/L, the protocol required a 25% dose reduction in the following cycles of chemotherapy. FN was defined as an ANC of less than 0.5 x 109/L, with a temperature  38.5°C (or developing a temperature of  38°C twice in a 12-hour period). According to current standard practice in the Netherlands, all patients with FN had to be admitted to the hospital and treated with parenteral broad-spectrum antibiotics (preferably a third-generation cephalosporin). In addition, standardized criteria had to be met before a patient could be discharged from the hospital, ie, temperature lower than 37.5°C for two consecutive days, no new clinical signs of infection, and ANC  0.5 x 109/L. Treating physicians were free to act in accordance with local hospital policy in treating patients with fever without neutropenia (ie, ANC  0.5 x 109/L), however cultures were mandatory. Prophylactic antibiotics were discontinued if therapeutic antibiotics had to be given, but G-CSF administration had to be continued.

    Statistical Analysis and Outcome Measures

    Primary outcome was defined as the difference between the two treatment groups in the proportion of patients with FN in the first cycle of chemotherapy. Secondary outcome measures were defined as the difference between the two prophylactic strategies in incidence of FN in cycles 2 to 5 and throughout the whole treatment period; incidence of fever in cycle 1, cycles 2 to 5, and in all cycles; number of clinically and microbiologically documented infections; cumulative dose and dose intensity. Additional explorative analyses were done comparing nadirs and duration of neutropenia between cycles.

    The primary parameter, the proportion of patients with FN during the first cycle of chemotherapy, was analyzed using the Cochran-Mantel-Haenszel test, combining results over nonempty strata. Categoric parameters were analyzed using the uncorrected 2 test. Continuous parameters were analyzed with Wilcoxon's two-sample test or the generalized Wilcoxon test when data were right censored. In the graphical presentation of nadir counts, medians and corresponding nonparametric 95% CIs are given.

    To investigate the influence of several possible predictors (or risk factors) on the occurrence of FN in the first cycle, logistic regression was used, both fixed and stepwise. The treatment parameter (G-CSF yes/no) was always forced in the models concerned. Candidate predictors were the four parameters that determined the 16 strata, all their first-order interactions and the four first-order interactions with the treatment parameter. The result of the stepwise procedure is presented graphically.

    All analyses were performed on the intention-to-treat population (defined as all randomized patients who started chemotherapy and signed informed consent). Only the hypothesis regarding the primary parameter was statistically tested with a two-sided level of significance of 5%. Before analysis, data of all centers were pooled. Likewise, data of all strata were pooled, except for analysis of the primary parameter.

    It should be noted that this study was powered (level of power 80%) to detect anticipated differences (15% reduction in incidence of FN in the first cycle, from 25% to 10%) in the primary parameter. Hence, both statistically significant (small P values) and nonsignificant (large P values) test results of secondary parameters and of additional exploratory analyses should be interpreted with caution. Apart from the primary parameter, no P values will be presented or referred to in the Results section. However, for ease of reference and to enable comparing our results with those published in relevant papers, P values are given in the tables.

    RESULTS

    Patients

    December 2000 until June 2003, 186 patients were recruited in 15 participating hospitals (one university medical center and 14 general hospitals) in the Netherlands. One hundred seventy-five patients were eligible (planned accrual 172 patients, including 10% ineligibility; Fig 1).

    Patient characteristics in both treatment arms were well balanced and are listed in Table 1. The majority of patients were older than 60 years, had ED, had a KS  80%, and received no previous chemotherapy.

    Febrile Neutropenia

    The addition of G-CSF to antibiotics prophylaxis significantly reduced the overall incidence of FN by nearly 50% in cycle 1; 20 patients (24%) in the antibiotics group developed FN compared with 9 patients (10%) in the antibiotics plus G-CSF group (P = .01; Table 2). The 20 patients with FN in the antibiotics group were similar to the nine patients in the antibiotics plus G-CSF group, with respect to stage, age, KS, and line of therapy.

    The incidence of FN was highest during cycle 1 and decreased thereafter in both arms. The incidence of FN was practically the same (17% v 11%) in both arms for cycles 2 to 5. But over the whole treatment period (cycle 1 to 5), the incidence of FN was significantly higher in the antibiotics arm (32% v 18%), but this distinction was largely due to the difference found in cycle 1.

    In total, 39 episodes of FN occurred in 27 of 85 patients in the antibiotics arm versus 21 episodes of FN in 16 of 90 patients in the antibiotics plus G-CSF arm. In cycles 2 to 5, FN occurred in 7% of 268 cycles in the antibiotics arm compared with 4% of 278 cycles in the antibiotics plus G-CSF arm. Although most patients experienced FN only once, FN itself (in any cycle) appeared to be a risk factor for developing FN in subsequent cycles, independent of the prophylactic strategy used. Approximately 50% of patients who continued chemotherapy in both arms developed at least one other episode of FN after an initial episode of FN (Table 3).

    In patients with FN during cycle 1, the duration of an episode of FN (median, 4 v 3 days), the duration of hospital admission for FN (median, 10 v 6 days), and the duration of therapeutic antibiotics needed because of FN (median, 7 v 8 days) differed marginally between both treatment groups. FN occurring during cycles 2 to 5 and over the whole treatment period was similar between the two treatment groups with regard to duration of FN, hospital admission, and therapeutic antibiotics (Table 2). In both study groups, 14 patients (17% and 16%, respectively) were admitted during cycle 1 for reasons other than FN (including fever/infection without neutropenia). The median duration among those patients who were hospitalized was longer in the antibiotics group (15 v 7 days). Regarding the entire treatment group, hospitalization due to all causes but FN was comparable between both study groups in cycle 1, (mean, 3.7 days v 1.8 days; means are presented here rather than medians because the majority of patients was not admitted and the median, therefore, is noninformative (zero in both groups). Duration of hospital admission (all causes) in cycle 1 was shorter in the antibiotics plus G-CSF group compared with the antibiotics-only group, due to the difference in incidence of FN (mean, 5.7 v 2.7 days).

    Eight patients died due to the sequels of FN; five from the antibiotics-only arm and three patients from the antibiotics plus G-CSF arm, ie, 19% of patients with FN in both arms, or 6% (arm A) and 3% (arm B) of the total study population. The majority of FN-related deaths occurred in the first cycle of treatment (4 in arm A and 2 in arm B) and all patients were older than 60 years (median, 69 years; range, 62 to 77 years).

    Delivered Chemotherapy and Tumor Response

    Sixty-two percent and 67% of patients completed the planned number of cycles in arms A and B, respectively. Reasons for discontinuation of chemotherapy were equally distributed in both treatment groups, the main reason being progressive disease (10% and 13%, respectively). The median delivered dose intensity of cyclophosphamide and doxorubicin for cycles actually delivered was statistically significantly higher in the antibiotics plus G-CSF arm, but the absolute difference was only 7.0 and 0.2 mg/m2/wk, respectively. The objective (complete and partial) response rate to chemotherapy was similar in both treatment arms (antibiotics-only group, 64%; antibiotics plus G-CSF group, 70%).

    Delivered Prophylaxis

    Antibiotics were started in 95% of delivered cycles (arm A + B) and G-CSF in 93% of delivered cycles (arm C). Prophylactic treatment with antibiotics and G-CSF was well tolerated, side effects being the reason for antibiotics discontinuation in only 5% (arm A) and 8% (arm B), and of G-CSF discontinuation in 2% of patients (Table 4).

    Toxicity

    The worst overall hematologic toxicity for all cycles is listed in Table 5. The addition of G-CSF not only lowered WBC counts and ANC nadir, but also shortened the median duration of WBC counts and ANC nadir in all cycles (Table 5). Of note, in both groups, the median ANC nadir was lower and the duration of grade 4 neutropenia was longer in cycle 1 compared with later cycles, however, in the antibiotics plus G-CSF group, this phenomenon is more pronounced (Fig 2; Table 5). Platelets nadir counts declined gradually during the whole treatment period, especially manifesting in the G-CSF arm (Fig 2). Treatment was delayed because of prolonged leuco-/neutropenia or complicated FN in 20 (26%) of 76 patients and three (4%) of 80 patients continuing chemotherapy after the first cycle in the antibiotics group and in the antibiotics plus G-CSF groups, respectively. But dose reduction for that reason was similar in both groups, that is, in 14 patients (18%) and nine patients (11%), respectively. Complicated FN was the main reason for the discontinuation of chemotherapy in three patients (one in antibiotics group), two patients (one in each group), zero patients, and two patients (one in each group) after cycles 1, 2, 3, and 4, respectively. Nonhematologic toxicities were mild and comparable between both study groups.

    Fever and Infections

    The incidence of fever in cycle 1 was higher in the antibiotics arm with an incidence of 27% in antibiotics-only arm versus 12% in the antibiotics plus G-CSF arm (Table 6). Over the whole treatment period, the difference in incidence of fever was of the same magnitude (40% v 28% of patients).

    The total number of infections appeared to be higher in the antibiotics arm (73 in 353 cycles or 0.21 infections per patient cycle v 50 in 368 cycles or 0.14 infections per patient cycle). But when excluding infections occurring during an episode of FN, there was barely any difference (0.096 v 0.079 infections per patient cycle). There was no relevant difference in clinically documented infections between both treatment groups over the whole treatment period, 44 in the antibiotics arm (60% of all infections) and 31 in the antibiotics plus G-CSF arm (62% of all infections). Culture-positive infections were more frequent in the antibiotics arm, especially in cycle 1 with 14 in the antibiotics arm (45% of all infections in cycle 1) versus three in the antibiotics plus G-CSF arm (17% of all infections in cycle 1). In patients with fever, the median duration of admission to the hospital in patients with fever was similar in both arms (11 v 14 days).

    Prognostic Factors of FN in the First Cycle

    The prophylactic strategy and the four parameters that defined the sixteen strata (elderly age, extensive stage of disease, poor KS, and second-line therapy) were considered risk factors for FN in the first cycle. First, a fixed logistic regression analysis was done with these five parameters as independent variables. Only the treatment parameter (odds ratio, 0.33 in cases where G-CSF was added) and age (dichotomized; odds ratio, 6.0 if  60 years) were independently related to the probability of FN in the first cycle (Table 7). To investigate whether some interactions were also related to this probability a stepwise logistic regression was done. As possible predictors beside the five main parameters above, all of their first-order interactions were chosen. In order to use as much information as possible, the continuous versions of age and KS were used instead of the dichotomous ones. Again, only the treatment parameter (odds ratio, 0.36; 95% CI, 0.15 to 0.84) and age (1.067 per year; 95% CI, 1.013 to 1.0124) were significantly related to the probability of FN in the first cycle (Fig 3).

    While planning the study we assumed that the number of risk factors would influence the risk of FN but this seemed not to be the case. The odds ratios for getting FN when patients had two or three to four risk factors instead were 1.55 (95% CI, 0.57 to 4.24) and 1.36 (95% CI, 0.44 to 4.19), respectively.

    DISCUSSION

    This randomized phase III study investigated the role of the addition of primary G-CSF prophylaxis to primary antibiotic prophylaxis in SCLC patients who were at risk of FN because of either elderly age, poor performance status, ED, and/or previous chemotherapy treatment. We showed that with the addition of primary G-CSF prophylaxis, the incidence of FN was further reduced in this high-risk population by more than 50% in the first chemotherapy cycle (24% v 10%).

    The incidence of FN in the first cycle in the antibiotics group is considerably greater than in the antibiotics arm of the previously published European Organisation for Research and Treatment of Cancer (EORTC) study by Tjan-Heijnen et al,2 and the incidence of FN in the first cycle is essentially the same as in the placebo group of that study, reflecting the (intended) selection of patients with a higher risk of FN in the present study. In the EORTC study, we selected patients with a good performance status to undergo intensified chemotherapy. In the present study, we selected patients known to be at risk of FN, that is, among others patients with a poor performance status and/or higher age. Based on the current study, only age of 60 years or older can be considered a risk factor. But considering the evidence from other studies, stage of disease and poor performance status cannot be rejected as risk factors.2,3,9 A systematic review of publications on risk models for FN has recently been published.16

    The incidence of FN in the second and subsequent cycles abruptly declined to 5% to 10% per cycle, with no clear differences between the two prophylactic treatment arms. The sample size used does not allow the detection of relatively small differences in secondary and additional exploratory parameters. Furthermore, we are aware of the risk of an increasing number of false-positive (type I errors) and false-negative (type II errors) results when the number of statistical tests involved increases. However, we feel confident that from all such results combined emerges a general picture of the processes involved.

    Although counterintuitive, others also observed that the risk of (complicated) neutropenia is greatest in the earliest cycles.2 In advanced breast cancer, 75% of observed episodes of FN occurred in the first cycle, and in non-Hodgkin's lymphoma, 63% of all toxic deaths occurred in cycle 1 (82% were infection related).17,18 The declining incidence of FN after the first cycle may reflect patient selection during treatment, as patients are more likely to continue treatment if suffering little chemotherapy-induced toxicity or having a tumor response. But in our study, 89% of patients received the second chemotherapy cycle in either prophylaxis arm and only three patients discontinued chemotherapy because of complicated FN. Also, we did not observe a difference in median chemotherapy dose between cycle 1 and subsequent cycles, which could have been another explanation of the declining incidence of FN (data not shown).

    Another explanation for the declining incidence of FN might be the lowered ANC nadir or the declining median duration of grade 4 neutropenia in cycles 2 to 5 compared with cycle 1. Only, this mainly occurred in the antibiotics plus G-CSF treatment group and cannot explain declining incidence in the antibiotics group. The lowered ANC nadir in subsequent cycles in the antibiotics plus G-CSF group suggests that there might be a carry over or priming effect of G-CSF administered in the first cycle to subsequent cycles and administration of G-CSF early in the course of treatment might be important. Priming is based on the idea of expanding the progenitor pool before the administration of cytotoxic medication, on the condition that G-CSF is withdrawn in time.19-23 The duration of neutropenia is also related to the occurrence of FN as is demonstrated in patients with acute myelogenous leukemia, breast cancer, urogenital cancer, and SCLC.18,24-26 Although in our patients in cycle 1 a longer duration of neutropenia was not associated with a higher probability of FN (logistic regression model, data not shown), others demonstrated in SCLC patients that each 1-day increase in duration of grade 4 neutropenia led to a 1.7-fold increase in the odds of developing FN.27 However, the observed declining duration of grade 4 neutropenia after the first cycle is more explicit in the antibiotics plus G-CSF group (Table 5) and therefore is unlikely to be responsible for the declining incidence of FN in subsequent cycles in both study groups. FN itself appeared to be a risk factor, as approximately 50% of patients receiving subsequent cycles of chemotherapy developed at least one other episode of FN. Thus, patients developing FN despite prophylaxis stay at risk in subsequent cycles despite continuing prophylactic measures. Whether an additional prophylactic measure can reduce the risk remains to be seen. The value of secondary prophylaxis is not well established. The current ASCO guideline on the use of secondary G-CSF prophylaxis is primarily based on data derived from the G-CSF registration trial in SCLC.1,10 That placebo-controlled trial allowed patients in the control arm who developed FN to receive open-label G-CSF in subsequent cycles of chemotherapy. The patients who were subsequently treated with open-label CSF had a reduction in the rate of FN from 100% in cycle 1 to 23% in cycle 2, despite receiving the same doses of chemotherapy. Because, as discussed before, the incidence of FN is highest in early cycles, it's questionable whether the addition of G-CSF is solely responsible for this decline.

    The addition of G-CSF to antibiotic prophylaxis did not influence FN-related mortality. Despite applying a combined prophylaxis, the FN-related mortality is still considerable (6% and 3%, respectively). One explanation might be that the patient population studied is not only at risk of developing FN but is also prone to develop complications of FN due to age, burden of disease, and comorbidity.16 Other investigators report similar mortality rates in similar patient populations.28,29 SCLC patients treated with CDE chemotherapy offer an opportunity to study the risk of FN as the risk of FN is increased by both the underlying disease and the use of the myelosuppressive chemotherapy regimen. An alternative strategy to prevent FN is to choose a less myelosuppressive chemotherapy regimen, as is the frequently used cisplatin-etoposide regimen. However, the use of cisplatin is limited by nephro- and neurotoxicity and use of hyperhydration, especially in the elderly. Surveys in the United Kingdom14 and in the Netherlands30 demonstrated that the CDE chemotherapy regimen is one of the most commonly used regimens for SCLC patients who are not candidates for combined chemoradiotherapy. A majority of Dutch respondents (75%), all pulmonologists with a special interest in lung cancer, prescribed CDE chemotherapy as first-line treatment for patients with SCLC. In the United Kingdom, CDE is the most common regimen administered in 28% of SCLC patients of any stage with a WHO performance status of 0 to 2 compared with the second most prescribed regimen (ifosfamide, carboplatin, etoposide) in 13%.

    Currently, in cancer treatment and support numerous costly treatment options are available and resources are only limited. Therefore, one cannot recommend an expensive measure without considering the costs. An economic evaluation was part of our study and provides stochastic cost data derived from the case report forms. Ensuing economic analyses will be described elsewhere in detail. However, the clinical data provide information sufficient for a rough calculation. The main cost driver in FN-related costs is hospitalization.31 In cycle 1, the total duration of FN-related hospitalization is 213 days for 85 patients and 87 days for 90 patients. In the Netherlands, the guideline price for a day of hospitalization is 270 (US $330).32 This means that for the first cycle, on average 1.5 days of hospitalization are saved, which is 405 (US $495) saved per patient, as opposed to the additional costs of G-CSF of 1,500 (US $1,833) per patient. A similar calculation for the entire treatment period results in an average savings of 2.1 days of hospitalization, or 567 (US $693) per patient saved, whereas the addition of five cycles of G-CSF costs an extra 7,500 (US $9,163). So, it can be anticipated that the addition of G-CSF is not cost saving.

    In conclusion, this study demonstrated that primary G-CSF prophylaxis added to primary antibiotic prophylaxis resulted in a statistically significant reduction in the incidence of FN in the first chemotherapy cycle. In particular, patients older than 60 years can benefit from the protective effects of G-CSF, because they have an increasing risk of FN. Although the combination results in a clinical benefit, a definite recommendation about whether to use or not to use the combination requires integration of clinical, quality of life, and economic data.

    Appendix

    Participating hospitals and responsible investigators: Rijnstate Ziekenhuis- Arnhem (H.J. Smit); Jeroen Bosch Ziekenhuis, GZG-'s-Hertogenbosch (B. Biesma); Ziekenhuis Gelderse Vallei-Ede (F.A. Wilschut); Radboud Universiteit Nijmegen Medisch Centrum-Nijmegen (J. Timmer-Bonte, V. Tjan-Heijnen, J. Festen, G. Bootsma); Streekziekenhuis Koningin Beatrix-Winterswijk (S. Cheragwandi); Canisius Wilhelmina Ziekenhuis-Nijmegen (A. Termeer); Diaconessenhuis-Meppel (C. Hensing); Medisch Centrum Rijnmond-Zuid, Zuiderziekenhuis Rotterdam (R. Slingerland); St Franciscus Gasthuis-Rotterdam (K. Tan); St Anna Ziekenhuis-Geldrop (C. vd Moosdijk); Slingeland Ziekenhuis-Doetinchem (G. Bosman); Isala klinieken, Weezenlanden-Zwolle (J. vd Berg); Maasziekenhuis-Boxmeer (R. Bunnik), the Netherlands.

    Authors' Disclosures of Potential Conflicts of Interest

    The authors indicated no potential conflicts of interest.

    Acknowledgment

    We thank Wim Lemmens, Department of Epidemiology and Biostatistics, Radboud Universiteit Nijmegen, the Netherlands, for his assistance in data processing and analysis and Frank van Leeuwen, head of Trial Office CCCE, Nijmegen, the Netherlands, for his advice on data management.

    NOTES

    Supported by a research grant from the Dutch Healthcare Insurance Board.

    Presented in part at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004.

    Authors' disclosures of potential conflicts of interest are found at the end of this article.

    REFERENCES

    Crawford J, Ozer H, Stoller R, et al: Reduction by granulocyte colony-stimulating factor of fever and neutropenia induced by chemotherapy in patients with small-cell lung cancer. N Engl J Med 325:164-170, 1991

    Tjan-Heijnen VC, Postmus PE, Ardizzoni A, et al: Reduction of chemotherapy-induced febrile leucopenia by prophylactic use of ciprofloxacin and roxithromycin in small-cell lung cancer patients: An EORTC double-blind placebo-controlled phase III study. Ann Oncol 12:1359-1368, 2001

    Morittu L, Earl HM, Souhami RL, et al: Patients at risk of chemotherapy-associated toxicity in small cell lung cancer. Br J Cancer 59:801-804, 1989

    Crawford J, Dale DC, Lyman GH: Chemotherapy-induced neutropenia: Risks, consequences, and new directions for its management. Cancer 100:228-237, 2004

    Engels EA, Ellis CA, Supran SE, et al: Early infection in bone marrow transplantation: Quantitative study of clinical factors that affect risk. Clin Infect Dis 28:256-266, 1999

    Cruciani M, Malena M, Bosco O, et al: Reappraisal with meta-analysis of the addition of gram-positive prophylaxis to fluoroquinolone in neutropenic patients. J Clin Oncol 21:4127-4137, 2003

    Tjan-Heijnen VC, Caleo S, Postmus PE, et al: Economic evaluation of antibiotic prophylaxis in small-cell lung cancer patients receiving chemotherapy: An EORTC double-blind placebo-controlled phase III study (08923). Ann Oncol 14:248-257, 2003

    Hughes WT, Armstrong D, Bodey GP, et al: 2002 guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 34:730-751, 2002

    Trillet-Lenoir V, Green J, Manegold C, et al: Recombinant granulocyte colony stimulating factor reduces the infectious complications of cytotoxic chemotherapy. Eur J Cancer 29:319-324, 1993

    Ozer H, Armitage JO, Bennett CL, et al: 2000 update of recommendations for the use of hematopoietic colony-stimulating factors: Evidence-based, clinical practice guidelines—American Society of Clinical Oncology Growth Factors Expert Panel. J Clin Oncol 18:3558-3585, 2000

    Lyman GH, Lyman CG, Sanderson RA, et al: Decision analysis of hematopoietic growth factor use in patients receiving cancer chemotherapy. J Natl Cancer Inst 85:488-493, 1993

    Bennett CL, Smith TJ, George SL, et al: Free-riding and the prisoner's dilemma: Problems in funding economic analyses of phase III cancer clinical trials. J Clin Oncol 13:2457-2463, 1995

    Adams JR, Lyman GH, Djubegovic B, et al: G-CSF as prophylaxis of febrile neutropenia in SCLC. Expert Opin Pharmacother 3:1273-1281, 2002

    Sambrook RJ, Girling DJ: A national survey of the chemotherapy regimens used to treat small cell lung cancer (SCLC) in the United Kingdom. Br J Cancer 84:1447-1452, 2001

    Thatcher N, Girling DJ, Hopwood P, et al: Improving survival without reducing quality of life in small-cell lung cancer patients by increasing the dose-intensity of chemotherapy with granulocyte colony-stimulating factor support: Results of a British Medical Research Council Multicenter Randomized Trial—Medical Research Council Lung Cancer Working Party. J Clin Oncol 18:395-404, 2000

    Lyman GH: Balancing the benefits and costs of colony-stimulating factors: A current perspective. Semin Oncol 30:10-17, 2003

    Gomez H, Hidalgo M, Casanova L, et al: Risk factors for treatment-related death in elderly patients with aggressive non-Hodgkin's lymphoma: Results of a multivariate analysis. J Clin Oncol 16:2065-2069, 1998

    Meza L, Baselga J, Holmes FA, et al: Incidence of febrile neutropenia is directly related to duration of severe neutropenia after myelosuppressive chemotherapy. Proc Am Soc Clin Oncol 21:255b, 2002 (abstr 2840)

    Broxmeyer HE, Benninger L, Patel SR, et al: Kinetic response of human marrow myeloid progenitor cells to in vivo treatment of patients with granulocyte colony-stimulating factor is different from the response to treatment with granulocyte-macrophage colony-stimulating factor. Exp Hematol 22:100-102, 1994

    Tjan-Heijnen VC, Biesma B, Festen J, et al: Enhanced myelotoxicity due to granulocyte colony-stimulating factor administration until 48 hours before the next chemotherapy course in patients with small-cell lung carcinoma. J Clin Oncol 16:2708-2714, 1998

    de Wit R, Verweij J, Bontenbal M, et al: Adverse effect on bone marrow protection of prechemotherapy granulocyte colony-stimulating factor support. J Natl Cancer Inst 88:1393-1398, 1996

    Hansen PB, Johnsen HE, Ralfkiaer E, et al: Short-term rhG-CSF priming before chemotherapy does mobilize blood progenitors but does not prevent chemotherapy induced myelotoxicity: A randomized study of patients with non-Hodgkin's lymphomas. Leuk Lymphoma 19:453-460, 1995

    Timmer-Bonte JNH, de Mulder PH, Peer PGM, et al: Timely withdrawal of G-CSF reduces the occurence of thrombocytopenia in dose-dense chemotherapy. Breast Cancer Res Treatment 93:117-123, 2005

    Heil G, Hoelzer D, Sanz MA, et al: A randomized, double-blind, placebo-controlled, phase III study of filgrastim in remission induction and consolidation therapy for adults with de novo acute myeloid leukemia: The International Acute Myeloid Leukemia Study Group. Blood 90:4710-4718, 1997

    Matsumoto T, Takahashi K, Tanaka M, et al: Infectious complications of combination anticancer chemotherapy for urogenital cancers. Int Urol Nephrol 31:7-14, 1999

    Meza LA, Green MD, Hackett JR, et al: Filgrastim-mediated neutrophil recovery in patients with breast cancer treated with docetaxel and doxorubicin. Pharmacotherapy 23:1424-1431, 2003

    Blackwell S, Crawford J: Filgrastim in the chemotherapy setting, in Morstyn G (ed): Filgrastim in Clinical Practice. New York, NY, Marcel Dekker, 1994, pp 103-116

    Klastersky J, Paesmans M, Rubenstein EB, et al: The Multinational Association for Supportive Care in Cancer risk index: A multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 18:3038-3051, 2000

    Talcott JA, Siegel RD, Finberg R, et al: Risk assessment in cancer patients with fever and neutropenia: A prospective, two-center validation of a prediction rule. J Clin Oncol 10:316-322, 1992

    Timmer-Bonte JNH, Tjan-Heijnen VC: Prevention of chemotherapy-related episodes of febrile neutropenia (FN) in small-cell-lung-cancer (SCLC) patients: In practice not theory. Eur J Cancer 39:S279, 2003 (abstr, suppl 5)

    Lyman GH, Kuderer NM: The economics of the colony-stimulating factors in the prevention and treatment of febrile neutropenia. Crit Rev Oncol Hematol 50:129-146, 2004

    Oostenbrink JB, Koopmanschap MA, Rutten FFH: Cost research manual, methods and guideline prices for economic evaluation in health care [Dutch]. Amstelveen: College voor Zorgverzekeringen, 2000

查询更多Chemotherapy-Induced相关信息在本站>>

《临床肿瘤学医学期刊》2005年11月第23卷第11期