J Clin Oncol 22: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 22 NUMBER 21 NOVEMBER JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Incidence and Predictors of Low Chemotherapy Dose-Intensity in Aggressive Non-Hodgkin s Lymphoma: A Nationwide Study Gary H. Lyman, David C. Dale, Jonathan Friedberg, Jeffrey Crawford, and Richard I. Fisher From the University of Rochester, Rochester, NY; University of Washington, Seattle, WA; and Duke University, Durham, NC. Submitted March 29, 2004; accepted August 18, Authors disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Gary H. Lyman, MD, MPH, James P. Wilmot Cancer Center, University of Rochester Medical Center, 601 Elmwood Ave, Rochester, NY 14642; Gary_Lyman@urmc.rochester.edu by American Society of Clinical Oncology X/04/ /$20.00 DOI: /JCO A B S T R A C T Purpose To assess the incidence of and risk factors for reduced relative dose-intensity (RDI) in patients treated with chemotherapy for aggressive non-hodgkin s lymphoma (NHL). Methods A nationwide survey was conducted of 567 oncology practices with data extracted from the records of 4,522 patients with aggressive NHL treated with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP); CHOP-rituximab (CHOP-R); or cyclophosphamide, mitoxantrone, vincristine, and prednisone (CNOP). The primary outcome was the average RDI for each regimen based on both planned and reference standards. Other assessments included the incidence of febrile neutropenia and patterns of colony-stimulating factor (CSF) use, as well as the average RDI in high-risk subgroups. Results Dose reductions 15% occurred in 40% of patients and treatment delays 7 days occurred in 24% of patients, resulting in 53% and 48% of patients receiving an RDI less than 85% of the minimum six-cycle and National Comprehensive Cancer Network guideline standards, respectively. Reduced RDI was more prevalent in older patients, with 60% of patients older than 60 years receiving RDI less than 85%. Multivariate analysis identified several independent predictors for reduced RDI, including age older than 60 years, advanced disease stage, poor performance status, and no prophylactic CSF use. Age was no longer a significant risk factor in patients who received prophylactic CSF. Conclusion Patients with aggressive and potentially curable NHL treated with CHOP, CHOP-R, or CNOP frequently receive reduced RDI. Predictive models based on the risk factors identified for reduced RDI should enable the targeted use of appropriate supportive care, facilitating the delivery of full chemotherapy doses on schedule. J Clin Oncol 22: by American Society of Clinical Oncology INTRODUCTION Malignant lymphoma, including both Hodgkin s disease and the non-hodgkin s lymphomas (NHLs), afflicts more than 50,000 individuals annually in the United States, resulting in nearly 20,000 deaths. 1 Combination therapy with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) is considered a standard treatment for patients with aggressive NHL. 2 Variations of the CHOP regimen include the addition of rituximab (CHOP- R) 3 and the substitution of mitoxantrone for doxorubicin (CNOP). 4 Adding rituximab to standard CHOP recently has been shown to be more effective than CHOP alone in elderly patients with newly diagnosed diffuse large B-cell lymphoma. 3 CNOP is frequently prescribed for older 4302

2 Low Chemotherapy Dose Intensity in NHL patients with NHL because mitoxantrone is perceived to be better tolerated and less cardiotoxic than doxorubicin. 5 Two large studies that compared CHOP and CNOP in the treatment of elderly patients with aggressive NHL, however, found that CHOP produced higher complete response rates and longer overall survival, without significantly greater toxicity. 5,6 The major dose-limiting toxicity with each of these regimens is myelosuppression and associated neutropenia, potentially compromising overall dose-intensity and long-term outcome The risk of neutropenic complications and the frequency of dose reductions and treatment delays in patients with NHL treated in randomized controlled trials have received some attention, but the treatment patterns in oncology practice remain largely unknown. 11 An analysis of data from patients with early-stage breast cancer has demonstrated that patients treated with adjuvant chemotherapy are at a substantial risk of chemotherapy dose attenuation, with more than half of all patients and two thirds of elderly patients being treated with a reduced chemotherapy dose-intensity. 12 This investigation was conducted to study patterns of chemotherapy treatment, including administered relative dose-intensity (RDI), as well as to identify risk factors for reduced RDI in patients with aggressive and potentially curable NHL. This is the largest study to date to evaluate practice patterns and treatment complications in patients with NHL treated with systemic chemotherapy. METHODS Study Design and Patient Selection Data were collected retrospectively between January 1, 1999, and December 31, 2001, on 6,314 patients with NHL who had been treated between 1993 and 2001 with chemotherapy at 567 oncology practices throughout the United States. Sites were instructed to obtain appropriate local institutional review board approval and were encouraged to use experienced oncology nurses, pharmacy personnel, or data management staff to collect the data. Patients were assigned unique numeric codes, with their identity known only to the individual sites. Sites were asked to identify and record data on 10 patients with NHL treated at the site, and reimbursement was based on the number of patient data forms submitted. The data were provided for analysis to the authors, who were blinded to the identities of the patients. Patients were included in the analysis if they had aggressive NHL (working formulation D to H) 13 and were treated with CHOP, CHOP-R, or CNOP. Only summary results are presented, with no breakdown by individual patient or treatment site. Practice sites were reimbursed by Amgen Inc (Thousand Oaks, CA) for their participation in this study. Chemotherapy Regimens and Study Outcomes The primary end point of the study was the average RDI (ARDI) for each regimen, defined as the proportion of the standard dose-intensity actually delivered. Although a range of RDI cut points were evaluated, for consistency with our previous report 12 the proportion of patients treated with less than 85% of standard was evaluated further. Reduced dose-intensity had two components: planned reductions in dose-intensity when the treatment was initiated and unplanned dose reductions and treatment delays that occurred during the course of therapy. Secondary end points included chemotherapy dose delays 7 days, dose reductions 15%, febrile neutropenia (FN) occurrence, and patterns of colony-stimulating factor (CSF) use. The dose and schedule of the CHOP, CHOP-R, and CNOP chemotherapy considered standard in this analysis are listed in Table 1. All regimens were administered in 21-day cycles. The agent-specific RDI in each patient was calculated as the ratio of the dose actually delivered over time to the standard dose-intensity. The ARDI was calculated by averaging the delivered RDIs of cyclophosphamide and doxorubicin for CHOP and CHOP-R, and by averaging the delivered RDIs of cyclophosphamide and mitoxantrone for CNOP. The following analyses were performed to address the stability of the RDI estimates under different clinical practice conditions and treatment policies: all patients according to the reference standard of a minimum of six cycles, and all patients according to the current National Comprehensive Cancer Drug Table 1. Chemotherapy Regimen Reference Standard Doses and Schedules CHOP (n 3,871) Regimen CHOP-R (n 117) CNOP (n 534) Doxorubicin 50 (day 1) 50 (day 1) Cyclophosphamide 750 (day 1) 750 (day 1) 750 (day 1) Vincristine 1.4; max 2 mg (day 1) 1.4; max 2 mg (day 1) 1.4; max 2 mg (day 1) Prednisone 100 mg po (days 1-5) 40 mg po (days 1-7) 100 mg po (days 1-5) Mitoxantrone 10 (day 1) Rituximab 375 (day 1) NOTE. All doses are in milligrams per square meter and are administered by intravenous infusion except as indicated. All regimens are given in 21-day cycles. The current guidelines of the NCCN 14 call for the administration of CHOP or CHOP-R for a minimum of six cycles in all high-risk patients, including those with stage II to IV disease and those with stage I disease and one or more adverse factors such as bulky adenopathy ( 10 cm), elevated LDH level, age older than 60 years, or ECOG performance status 2. For low-risk patients (stage I disease without any of these adverse factors), the guidelines call for a minimum of three cycles along with locoregional radiation therapy. Abbreviations: CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; CHOP-R, CHOP-rituximab; CNOP, cyclophosphamide, mitoxantrone, vincristine, and prednisone; po, orally; NCCN, National Comprehensive Cancer Network; LDH, lactate dehydrogenase; ECOG, Eastern Cooperative Oncology Group; ARDI, average relative dose-intensity. These drugs are not used in calculating the ARDI

3 Lyman et al Network (NCCN) recommendation of a minimum of six cycles in high-risk patients and three cycles along with regional radiation therapy in low-risk, low-stage patients 14 (Table 1). High-risk patients based on NCCN guidelines consist of those with stage II to IV disease and patients with stage I disease with one or more adverse risk factors, including age older than 60 years, Eastern Cooperative Oncology Group (ECOG) performance status 2, elevated serum lactate dehydrogenase (LDH), bulky adenopathy ( 10 cm), or at least two extranodal sites of disease. Low-risk patients were defined as those with stage I disease without any adverse risk factors. Subgroup analyses were limited to either high-risk patients according to the NCCN criteria or patients with stage III or IV disease according to a reference standard of a minimum of six cycles. Dose and schedule sufficient to calculate RDI was available in 4,513 (99.85%) patients, including all but nine high-risk patients and six stage III to IV patients. Study Independent Variables and Operational Definitions Patient demographics and clinical characteristics collected included age, sex, height, weight, body-surface area (BSA), chemotherapy regimen, and planned dose and schedule. In addition, disease histology by International Working Formulation, Ann Arbor stage, number of extranodal sites, presence of bone marrow involvement (where assessed), ECOG performance status, LDH, and serum albumin data were collected. Cycle-specific information included actual chemotherapy doses and interval and pretreatment WBC count or absolute neutrophil count (ANC). Episodes of FN were captured for each patient over all cycles, but not separately for each cycle. Age was dichotomized as 60 or older than 60 years because these patient groups have different risks of death according to the International Prognostic Index. 15 BSA was categorized as 2 or more than 2 m 2. Disease stage was categorized as limited stage (Ann Arbor stage I or II) or advanced stage (Ann Arbor stage III or IV). The number of extranodal sites was classified as less than 2 or 2. Serum LDH level was categorized as 500 or less than 500 U/L, and serum albumin level was classified as 3.5 or more than 3.5 g/dl. Lymphoma histology was originally categorized according to the International Working Formulation that was commonly used when the disease was diagnosed and the patients were treated. 13 The analysis reported here was limited to aggressive histology NHL (Working Formulation D to H). Patterns of CSF use were determined by start date, duration, and dose assessed overall and for each cycle of chemotherapy. Prophylactic CSF was defined as its use within 72 hours of chemotherapy. Primary prophylactic CSF was defined as prophylactic use beginning in the first cycle of chemotherapy. Treatment response (complete, partial, stable disease, progressive disease, or unknown) and cycle-specific FN rates were added as variables during the course of the study and are only available for a subgroup (n 1,400; 31%) of patients. Mortality was not reported in this study and the limited duration of follow-up does not permit any estimation of disease-free or overall survival as outcomes. Statistical Methods The distribution of each demographic and clinical variable was reviewed, and appropriate summary measures were estimated, including means, medians, and proportions. The relation of each demographic and clinical variable to the primary and secondary outcomes was evaluated in univariate analysis. Group comparisons were based on 2 distributions for categoric variables, and those for continuous variables were based on Student s t test for normally distributed variables and the Mann-Whitney U statistic for all other variables. For univariate comparisons and multivariate analyses, ARDI was categorized as less than 85% or 85%. Multivariate analysis was based on fixed logistic regression models for ARDI less than 85%, with the covariates specified in advance. Regression coefficients reflect the rate of change of the dependent (outcome) variable for each unit change in the independent variable (covariate) adjusted for all other variables in the model. Global model significance was based on the 2 method, and the significance of individual covariates was based on the Wald statistic. The exponential function of the regression coefficient for dichotomous variables represents the adjusted odds ratio (OR) as an estimate of the relative risk. The odds ratios presented are adjusted for any significant continuous or multichotomous variables in the model. Two-sided tests of the null hypothesis were used throughout. Models are presented without interaction terms. However, first-order interaction terms were evaluated for each model and significant interactions are reported. RESULTS Patient Characteristics and Chemotherapy Regimens Of the 6,314 eligible patients, 4,839 (77%) were classified as having aggressive NHL, 655 (10%) were classified as having indolent NHL, and 132 (2%) were classified as having highly aggressive NHL, whereas 688 (11%) were unclassifiable or unclassified. The analysis reported here is limited to the 4,522 patients with aggressive histology NHL who were treated with CHOP (n 3,871; 86%), CHOP-R (n 117; 3%), or CNOP (n 534; 11%). Patient characteristics are summarized in Table 2 on all patients (n 4,522), high-risk patients according to NCCN criteria (n 4,102), and patients with stage III or IV disease (n 2,493). Diffuse large-cell lymphoma (DLCL) was the single most common histology (56.8%). The mean age of the study population was 61.2 years (range, 16 to 96 years); 58.3% of patients were older than 60 years of age, and more than half of the patients (52.6%) were female. RDI Actual, planned, and reference chemotherapy doses were standardized to individual calculated BSA. Intention to deliver fewer than six cycles of chemotherapy was encountered in 30%, 28%, and 20% of all patients, NCCN high-risk patients, and patients with stage III or IV disease, respectively (Fig 1). As shown in Figure 2, the distributions of ARDI estimates were similar when compared with the different standards: Figure 2A, six cycles for all patients (median, 0.83); Figure 2B, based on NCCN guidelines (median, 0.86); Figure 2C, six cycles for high-risk patients based on NCCN guidelines (median, 0.83); and Figure 2D, six cycles for patients with stage III or IV disease (median, 0.86). Although some patients were treated with more than the reference standard dose-intensity, most received less. As shown in Figure 3, the incidences of dose reductions, 4304 JOURNAL OF CLINICAL ONCOLOGY

4 Low Chemotherapy Dose Intensity in NHL Characteristic Table 2. Clinical and Treatment Characteristics in Patients With Aggressive NHL All Patients (n 4,522) High Risk (n 4,102) Stage III or IV Disease (n 2,493) NCCN risk criteria, % Stage I Stage II Stage III Stage IV Age 60 years, % ECOG performance status 2, % Elevated LDH, % Extranodal involvement 2, % Any high risk, % International Prognostic Index, % Histology, % Diffuse large-cell Other Any CSF use, % Any prophylactic CSF, % Primary (1st-cycle) CSF use, % Abbreviations: NHL, non-hodgkin s lymphoma; NCCN, National Comprehensive Cancer Network; ECOG, Eastern Cooperative Oncology Group; LDH, lactate dehydrogenase; CSF, colony-stimulating factor. Stages II to IV, age older than 60 years, ECOG performance status 2, bulky disease 10 cm, elevated LDH. treatment delays, and RDI less than 85% were comparable regardless of which standard or patient criteria were analyzed, with anywhere from 48% to 53% of patients receiving less than 85% of standard dose-intensity overall. Significantly greater dose reductions and treatment delays were Fig 1. Planned number of cycles of chemotherapy in all non-hodgkin s lymphoma patients, high-risk patients based on National Comprehensive Cancer Network (NCCN) criteria, and patients with stage III or IV disease. observed in patients treated with CNOP than with CHOP or CHOP-R, with 62% of those treated with CNOP receiving less than 85% of the standard dose-intensity (P.001). As shown in Figure 4, a range of threshold reductions in dose-intensity were studied. Figure 5 shows the proportion of patients receiving less than 85% of standard dose-intensity during the first five cycles of treatment in all patients and by age group. A significant trend toward reduced dose-intensity was observed across cycles of chemotherapy (P.001). However, the greatest change in either planned or standard RDI was at the start of treatment in cycle 1. As shown in Figure 5, elderly patients consistently received lower dose-intensity than did younger patients. A greater proportion of older patients (older than 60 years) received an RDI less than 85% during the first five cycles of treatment, with 60% receiving RDI less than 85% overall compared to 44% of those younger than 60 years of age. Response to treatment was available only in a subgroup of patients (n 1,400). Reported responses included complete response (53%), partial response (17%), stable disease (4%), progressive disease (6%), and unknown (20%). The average RDI based on treatment response was complete response (0.79), partial response (0.82), stable disease (0.77), progressive disease (0.78), and unknown (0.75). The proportion of patients

5 Lyman et al Fig 2. Relative dose-intensity (RDI) in patients with aggressive non-hodgkin s lymphoma based on (A) a standard of six cycles for all patients, (B) a standard based on National Comprehensive Cancer Network (NCCN) guidelines for all patients, (C) a standard based on NCCN guidelines for high-risk patients, and (D) a standard based on NCCN guidelines for patients with stage III or IV disease. receiving an RDI less than 85% by response included complete response (51%), partial response (45%), stable disease (57%), progressive disease (56%), and unknown (59%). Fewer responding patients received RDI less than 85% than did nonresponding patients (P.021). When the actual dose-intensity was compared with that planned in all patients, the mean and median ARDI was and 0.939, respectively, with 32.4% of patients receiving less than 85% of the planned dose-intensity. Similar proportions of high-risk patients and of those with stage III and IV disease received less than 85% of the planned doseintensity (33.7% and 38.3%, respectively). The mean and median planned dose-intensity relative to standard six cycles of therapy in all patients was and 0.975, respectively, with RDI less than 85% planned in 37.9% of patients. Similar proportions of patients treated with planned RDI less than 85% of the standard were observed in all patients based on NCCN guidelines (33.3%), in high-risk patients (36.5%), and in patients with stage III or IV disease (30.2%). Of the overall average reduction in ARDI of 0.213, (47%) was planned from the start of therapy and (53%) was unplanned and associated with subsequent dose reductions and treatment delays. Figure 6 illustrates the planned and unplanned reductions along with actual ARDI for all patients relative to either a standard six cycles or based on NCCN guidelines as well as for high-risk patients and for patients with stage III or IV disease. Planned and unplanned reductions in dose-intensity were greater in patients older than 60 years (12.7% v 11.7%) compared with younger patients (11.7% v 10.7%). As shown in Figure 7, significant linear trends were observed during the study period for decreases in unplanned dose reductions (P.001) along with increases in planned dose reductions (P.001), with minimal change in the overall ARDI. Table 3 lists the adjusted ORs for RDI less than 85% based on multivariate models for the different reference standards considered. Significant predictors of reduced 4306 JOURNAL OF CLINICAL ONCOLOGY

6 Low Chemotherapy Dose Intensity in NHL Fig 3. Patients with treatment delays 7 days, dose reductions 15%, and relative dose-intensity (RDI) less than 85%, by treatment regimen based on standards of (A) six cycles for all patients, (B) National Comprehensive Cancer Network (NCCN) guidelines for all patients, (C) NCCN guidelines for high-risk patients, and (D) NCCN guidelines for patients with stage III or IV disease. CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; CHOP-R, CHOP-rituximab; CNOP, cyclophosphamide, mitoxantrone, vincristine, and prednisone. RDI that were common to most models were age older than 60 years, poor performance status, earlier year of treatment, and practice site, whereas prophylactic CSF was associated with significantly less reduction in dose-intensity. The odds ratios presented are adjusted for year of treatment, practice site, and planned number of cycles, which were also significant predictors in each of the models shown. Lower RDI was observed over time and was associated most notably with a decrease in planned RDI. Approximately half of the variance in reduced RDI was explained by the covariates in these models. Significant interaction was observed between CSF prophylaxis and age for reduced dose-intensity (P.049). Older age was a significant predictor of reduced dose-intensity in patients not treated with prophylactic CSF (OR, 2.41) but did not reach significance in patients treated with prophylactic CSF (OR, 1.42). Multivariate models based on a range of threshold reductions in dose-intensity provided essentially identical results to those shown in Table 3. FN Midcycle ANC less than /L occurred in 62% and ANC less than /L occurred in 35% of all patients. One or more episodes of FN occurred in 945 patients (20.9%), with hospitalization required in 717 (76%). The risk of FN was greatest in patients with age older than 60 years (24.8%), BSA 2m 2 (22.6%), ECOG performance status 2 (34.6%), stage III or IV disease (23.5%), International Prognostic Index 2 (26.7%), and treatment with CHOP-R (25.6%). In multivariate analysis, significant independent predictors of FN were older age, female sex, advanced disease stage, poor performance status, and low pretreatment ANC (Table 4). The odds ratios presented are adjusted for year of treatment and disease histology, which

7 Lyman et al Fig 4. Patients receiving reduced dose-intensity based on a standard of six cycles or National Comprehensive Cancer Network (NCCN) guidelines for all patients, high-risk patients based on NCCN criteria, and patients with stage III or IV disease. were also significant predictors of FN in this model. Lower risk of FN was observed over time and in those receiving primary CSF prophylaxis. A significant interaction was observed between CSF prophylaxis and age for risk of FN (P.001). Older age was a significant risk factor for FN when prophylactic CSF was not used (OR, 1.80) but not when prophylactic CSF was used (OR, 0.77). CSF Use CSF support was used during the course of chemotherapy in 2,444 patients (54%), including 1,325 (29.3%) during the first cycle. CSF was used as primary prophylaxis in 541 patients (12.0%), accounting for 22% of CSF use (Fig 8). First-cycle CSF use was initiated a mean of 7.2 days (median, 7.0 days) after chemotherapy, with CSF therapy started after an interval more than 5 days in 53% of patients. More than half of the patients were treated with CSF for less than 7 days. The use of prophylactic CSF increased during the first three cycles (Fig 8). The duration of CSF use also increased after the first cycle, whereas the interval between Fig 6. Average relative dose-intensity (ARDI) based on standard doseintensity for six cycles in all patients, National Comprehensive Cancer Network (NCCN) guidelines for all patients, NCCN guidelines for high-risk patients, and NCCN guidelines for patients with stage III or IV disease. the administration of chemotherapy and that of CSF decreased. In multivariate analysis, the use of prophylactic CSF starting in cycle 1 was significantly associated with a lower risk of FN (Table 4). The initiation of prophylactic CSF later than cycle 1 was not associated with a significant reduction in the risk of FN. In addition, when CSF was initiated prophylactically during the treatment course, multivariate analysis showed less reduction in RDI (Table 3). As noted in Methods, cycle-specific rates of FN were available only in a subgroup of patients (n 1,400). In patients not receiving CSFs during the first cycle, 10.2% without FN and 55.3% with FN went on to receive prophylactic CSF in cycle 2 (P.0001). Similar numbers held for patients experiencing ANC less than /L or less than /L during cycle 1 (P.0001). Likewise, significantly fewer patients experienced reduced dose-intensity in cycle 2 who received primary prophylaxis in cycle 1 (P.05). DISCUSSION This retrospective analysis of records on more than 4,500 patients with aggressive-histology NHL treated with CHOP Fig 5. Proportions of patients receiving relative dose-intensity less than 85% in the first five cycles of chemotherapy and overall. Results are shown for all patients, those age 60 years, and patients older than 60 years. Fig 7. Average relative dose-intensity (ARDI) along with planned and unplanned dose reductions, by year of treatment between 1995 and JOURNAL OF CLINICAL ONCOLOGY

8 Low Chemotherapy Dose Intensity in NHL Table 3. Multivariate Logistic Regression Analyses for Relative Dose-Intensity 85% in Patients With Aggressive NHL All Patients, 6 Cycles (n 4,513) All Patients, NCCN Criteria (n 4,513) High Risk, 6 Cycles (n 4,093) Stage III or IV, 6 Cycles (n 2,486) Covariate OR 95% CI OR 95% CI OR 95% CI OR 95% CI Age 60 years to to to to 2.85 Stage III to to to 1.42 ECOG performance status to to to to 1.77 Albumin 3.5 g/dl to to 1.82 Prophylactic G-CSF to to to to 0.94 R C statistic to to to to 0.79 Abbreviations: NHL, non-hodgkin s lymphoma; NCCN, National Comprehensive Cancer Network; ECOG, Eastern Cooperative Oncology Group; G-CSF, granulocyte colony-stimulating factor; OR, odds ratio. ORs adjusted for year of treatment; planned duration of treatment, and practice site. or CHOP-like chemotherapy in clinical practices throughout the United States found substantial reductions in chemotherapy dose-intensity in half of the patients. This large practice-based study has shown that both planned and unplanned chemotherapy dose modifications occur, resulting in reduced ARDI. A number of factors were found to be significantly associated with reduced dose-intensity in both univariate and multivariate analysis, including older age, poor performance status, advanced disease stage, year of treatment, and practice site. The age distribution of patients in this study reflects that of the broad population of patients in the United States with NHL: More than half are age 60 or older. 16 Given that advanced age is associated with a higher incidence of neutropenic complications, the delivery of standard dose-intensity may be particularly problematic in the elderly. The results reported here for aggressive NHL are similar to those recently observed in nearly 20,000 women receiving adjuvant breast cancer chemotherapy. 12 Although data on long-term outcomes are not available from these surveys, previous studies have raised concern about the impact of reduced dose-intensity on survival in patients with NHL Retrospective analysis of data on 115 patients with DLCL treated with chemotherapy found that an RDI of doxorubicin of more than 75% was the most important predictor of survival. 8 Epelbaum et al 7 reported a 5-year survival rate of 80% in patients with DLCL who were treated with more than 70% RDI of CHOP in their first cycle of chemotherapy and a 5-year survival rate of only 32% in those receiving less than 70% (P.0001). In a recent study of doxorubicin-based chemotherapy for DLCL, 5-year survival was lower in patients age 60 years than in younger patients (30% v 57%; P.001). 10 However, the subset of elderly patients in this study who were treated with doxorubicin at dose intensities 10 mg/ m 2 /wk had a 5-year survival rate similar to that in younger patients (52%). Recent studies in patients with aggressive NHL provide evidence for a possible advantage for increased chemotherapy dose-intensity based on administration every 14 days with granulocyte CSF support. In a study from the German High-Grade NHL Study Group, such dose-dense CHOP in elderly patients with NHL resulted in both longer time to treatment failure and improved overall survival than did Table 4. Multivariate Logistic Regression Analysis for Febrile Neutropenia in Patients With Aggressive NHL (N 4,522) Covariate Adjusted OR 95% CI Age 60 years to 1.84 Female sex to 1.55 Stage III to 1.51 ECOG performance status to 2.34 BSA 2m to 1.04 Pretreatment ANC /L to 2.21 Prophylactic CSF starting in the first cycle to 0.92 NOTE. R ; C statistic ( ). Abbreviations: NHL, non-hodgkin s lymphoma; OR, odds ratio; ECOG, Eastern Cooperative Oncology Group; BSA, body-surface area; ANC, absolute neutrophil count; CSF, colony-stimulating factor. ORs adjusted for year of treatment and histologic type. Fig 8. Patients receiving prophylactic and nonprophylactic colonystimulating factor, by cycle and overall

9 Lyman et al standard 21-day CHOP. 21 In a similar randomized controlled trial from the same group in younger patients with good-prognosis aggressive NHL treated with CHOP or CHOP-etoposide, 14-day schedules were associated with improved overall survival in univariate (P.050) and multivariate (P.044) analysis. 22 Randomized clinical trials of the CHOP, CHOP-R, and CNOP regimens have consistently reported that myelosuppression in general and neutropenia in particular represent the major dose-limiting toxicities. In our study, a 20% incidence of FN was associated with a high degree of dose attenuation. However, it is possible that some of the unplanned reductions in dose-intensity were due to nonhematologic toxicity. Our study lacks detailed information on the frequency of nonhematologic toxicity and its potential influence on reduced dose-intensity. Of particular concern is the apparent trend toward increasing reductions in planned dose-intensity over time, given that such reductions cannot be accounted for on the basis of myelosuppression or patient compliance issues. Although some reductions may be considered appropriate in the elderly and those with serious comorbidities, the planned reductions may also represent a systematic intent to undertreat patients with aggressive NHL compared with published reference standard regimens in some practices and settings. In addition, of some concern is the significant variation in RDI between practice sites, although limited data are available on comorbidities that might have resulted in reductions in delivered dose-intensity. Such data were collected only during the last 2 years of the survey and are not available on the majority of the patients in this study. It is of interest that practice site was significantly associated with planned but not unplanned decreases in RDI. Likewise, stage 3, ECOG 2, and albumin 3.5 g/dl were significantly associated with unplanned but not planned reductions, whereas age older than 60 years was associated with both. In the absence of data on disease-free and overall survival, the impact of reduced RDI and the CSFs could not be addressed in our study. A substantial proportion of patients either received no CSF support or received it only after a neutropenic complication. As reported, most patients receiving prophylactic CSF only after cycle 1 did so after experiencing either severe or febrile neutropenia during cycle 1. Randomized clinical trials have shown that prophylactic CSF can reduce the risk of neutropenic complications 6,23-25 and facilitate the delivery of full dose-intensity. 6,26-30 Although our study is not a randomized controlled trial, it is supportive of these trials in demonstrating that CSFs administered prophylactically starting with the first cycle are associated with a reduction in the risk of FN and reduced dose-intensity in multivariate analysis. A disturbing pattern of nonprophylactic use of the CSFs late in the treatment cycle was observed in this study. Such use of the CSFs has not clearly been demonstrated to be efficacious. Although prophylactic use appears to increase over reported cycles, much of the later use is secondary to previous severe and febrile neutropenia events. Two studies in patients with NHL treated with CHOP have shown that maintaining chemotherapy dose-intensity with growth factor support in both 21-day and dose-dense 14- day regimens was not associated with a higher incidence of nonhematologic toxicity. 31,32 In our study, several factors were associated with reduced dose-intensity in both univariate and multivariate analysis. Older patients in this study had more than twice the risk of lower dose-intensity CHOP chemotherapy than did younger patients, after adjustment for other risk factors. These data are consistent with those from a retrospective study of 577 patients with intermediate-grade NHL who were treated with CHOP, which showed that planned and actual chemotherapy dose intensities were significantly lower in older patients ( 65 years) than in younger patients. 33 Concern has been expressed about the possible relationship between the reduced RDI for chemotherapy in older patients with NHL and the poor prognosis reported. 15,34,35 Guidelines published by the NCCN recommend preemptive use of hematopoietic growth factors in patients aged 70 years and older who are treated with myelosuppressive chemotherapy similar in intensity to CHOP. 20 The multivariate analysis reported here suggests that, after adjustment for age and the other risk factors considered, the prophylactic use of CSF is associated with less reduction in RDI. Patients who began to receive CSF may have been at higher risk for FN and reduced dose-intensity than those not receiving CSF, potentially biasing the study against the effect of CSFs on these outcomes. Therefore, the significant impact of primary prophylaxis on FN risk and reduced dose-intensity after adjustment for other known risk factors is perhaps even more reassuring. Prospective trials of the effect of targeted supportive care on RDI and treatment outcomes, including disease-free and overall survival in patients with NHL, are needed. Authors Disclosures of Potential Conflicts of Interest The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Performed contract work within the last 2 years: Gary H. Lyman, Amgen, Glaxo. Received more than $2,000 a year from a company for either of the last 2 years: Gary H. Lyman, Amgen Speaker s Bureau, Orthobiotech Speaker s Bureau; Jeffrey Crawford, Amgen Speaker s Bureau; David C. Dale, Amgen Speaker s Bureau JOURNAL OF CLINICAL ONCOLOGY

10 Low Chemotherapy Dose Intensity in NHL REFERENCES 1. Jemal A, Tiwari RC, Murray T, et al: Cancer statistics, CA Cancer J Clin 54:8-29, Fisher RI, Gaynor ER, Dahlberg S, et al: Comparison of a standard regimen (CHOP) with three intensive chemotherapy regimens for advanced non-hodgkin s lymphoma. N Engl J Med 328: , Coiffier B, Lepage E, Brière J, et al: CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large B-cell lymphoma. N Engl J Med 346: , Pavlovsky S, Santarelli MT, Erazo A, et al: Results of a randomized study of previouslyuntreated intermediate and high grade lymphoma using CHOP versus CNOP. Ann Oncol 3: , Sonneveld P, de Ridder M, van der Lelie H, et al: Comparison of doxorubicin and mitoxantrone in the treatment of elderly patients with advanced diffuse non-hodgkin s lymphoma using CHOP versus CNOP chemotherapy. J Clin Oncol 13: , Osby E, Hagberg H, Kvaloy S, et al: CHOP is superior to CNOP in elderly patients with aggressive lymphoma while outcome is unaffected by filgrastim treatment: Results of a Nordic Lymphoma Group randomized trial. Blood 101: , Epelbaum R, Haim N, Ben-Shahar M, et al: Dose intensity analysis for CHOP chemotherapy in diffuse aggressive large cell lymphoma. Isr J Med Sci 24: , Kwak LW, Halpern J, Olshen RA, et al: Prognostic significance of actual dose intensity in diffuse large-cell lymphoma: Results of a treestructured survival analysis. J Clin Oncol 8: , Lepage E, Gisselbrecht C, Haioun C, et al: Prognostic significance of received relative dose intensity in non-hodgkin s lymphoma patients: Application to LNH-87 protocol: The GELA (Groupe d Etude des Lymphomes de l Adulte). Ann Oncol 4: , Lee KW, Kim DY, Yun T, et al: Doxorubicinbased chemotherapy for diffuse large B-cell lymphoma in elderly patients: Comparison of treatment outcomes between young and elderly patients and the significance of doxorubicin dosage. Cancer 98: , Dale DC, McCarter G, Crawford J, et al: Myelotoxicity and dose intensity of chemotherapy: Reporting practices from randomized clinical trials. J Natl Comp Cancer Network 1: , Lyman GH, Dale DC, Crawford J: Incidence and predictors of low dose-intensity in adjuvant breast cancer chemotherapy: A nationwide study of community practices. J Clin Oncol 21: , Non-Hodgkin s Lymphoma Pathologic Classification Project: National Cancer Institute sponsored study of classifications of non- Hodgkin s lymphomas: Summary and description of a working formulation for clinical usage. Cancer 49: , National Comprehensive Cancer Network: NCCN Clinical Practice Guidelines in Oncology: Non-Hodgkin s Lymphoma, Version 1, nhl.pdf 15. International Non-Hodgkin s Lymphoma Prognostic Factors Project: A predictive model for aggressive non-hodgkin s lymphoma. N Engl J Med 329: , Ries LAG, Eisner MP, Kosary CL, et al: SEER Cancer Statistics Review, Armitage JO, Potter JF: Aggressive chemotherapy for diffuse histiocytic lymphoma in the elderly: Increased complications with advancing age. J Am Geriatr Soc 32: , Gomez H, Mas L, Casanova L, et al: Elderly patients with aggressive non-hodgkin s lymphoma treated with CHOP chemotherapy plus granulocyte-macrophage colony-stimulating factor: Identification of two age subgroups with differing hematologic toxicity. J Clin Oncol 16: , Morrison VA, Picozzi V, Scott S, et al: The impact of age on delivered dose intensity and hospitalizations for febrile neutropenia in patients with intermediate-grade non-hodgkin s lymphoma receiving initial CHOP chemotherapy: A risk factor analysis. Clin Lymphoma 2:47-56, Balducci L, Yates J: General guidelines for the management of older patients with cancer. Oncology (Huntingt) 14: , Pfreundschuh M, Truemper L, Kloess M, et al: 2-weekly or 3-weekly CHOP chemotherapy with or without etoposide for the treatment of elderly patients with aggressive lymphomas: Results of the NHL-B2 trial of the DSHNHL. Blood 104: , Pfreundschuh M, Truemper L, Kloess M, et al: 2-weekly or 3-weekly CHOP chemotherapy with or without etoposide for the treatment of young patients with good prognosis (normal LDH) aggressive lymphomas: Results of the NHL-B1 trial of the DSHNHL. Blood 104: , 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: , Trillet-Lenoir V, Green J, Manegold C, et al: Recombinant granulocyte colony stimulating factor reduces the infectious complications of cytotoxic chemotherapy. Eur J Cancer 29A: , Lyman GH, Kuderer NM, Djulbegovic B: Prophylactic granulocyte colony-stimulating factor in patients receiving dose-intensive cancer chemotherapy: A meta-analysis. Am J Med 112: , Bertini M, Freilone R, Vitolo U, et al: P- VEBEC: A new 8-weekly schedule with or without rg-csf for elderly patients with aggressive non-hodgkin s lymphoma (NHL). Ann Oncol 5: , Doorduijn JK, van der Holt B, van Imhoff GW, et al: CHOP compared with CHOP plus granulocyte colony-stimulating factor in elderly patients with aggressive non-hodgkin s lymphoma. J Clin Oncol 21: , Pettengell R, Gurney H, Radford JA, et al: Granulocyte colony-stimulating factor to prevent dose-limiting neutropenia in non-hodgkin s lymphoma: A randomized controlled trial. Blood 80: , Zagonel V, Babare R, Merola MC, et al: Cost-benefit of granulocyte colony-stimulating factor administration in older patients with non- Hodgkin s lymphoma treated with combination chemotherapy. Ann Oncol 5: , 1994 (suppl 2) 30. Zinzani PL, Pavone E, Storti S, et al: Randomized trial with or without granulocyte colonystimulating factor as adjunct to induction VNCOP-B treatment of elderly high-grade non- Hodgkin s lymphoma. Blood 89: , Jacobson JO, Grossbard M, Shulman LN, et al: CHOP chemotherapy with preemptive granulocyte colony-stimulating factor in elderly patients with aggressive non-hodgkin s lymphoma: A doseintensity analysis. Clin Lymphoma 1: , Wunderlich A, Kloess M, Reiser M, et al: Practicability and acute haematological toxicity of 2- and 3-weekly CHOP and CHOEP chemotherapy for aggressive non-hodgkin s lymphoma: Results from the NHL-B trial of the German High-Grade Non-Hodgkin s Lymphoma Study Group (DSHNHL). Ann Oncol 14: , Lyman GH, Morrison VA, Dale DC, et al: Risk of febrile neutropenia among patients with intermediate-grade non-hodgkin s lymphoma receiving CHOP chemotherapy. Leuk Lymphoma 44: , Dixon DO, Neilan B, Jones SE, et al: Effect of age on therapeutic outcome in advanced diffuse histiocytic lymphoma: The Southwest Oncology Group experience. J Clin Oncol 4: , Maartense E, Kluin-Nelemans HC, Noordijk EM: Non-Hodgkin s lymphoma in the elderly: A review with emphasis on elderly patients, geriatric assessment, and future perspectives. Ann Hematol 82: ,