Dose intensity and hematologic toxicity in older breast cancer patients receiving systemic chemotherapy

Abstract

BACKGROUND:

This prospective study evaluated patient and treatment characteristics that contributed to hematologic toxicity in older breast cancer patients treated with curative intent in the community setting.

METHODS:

Data were collected on 1224 patients with stage I through III breast cancer, of whom 207 were aged ≥65 years (grading determined according to the American Joint Committee on Cancer staging system). Primary outcome measures included anemia, thrombocytopenia, severe neutropenia, febrile neutropenia, and both planned and actual relative dose intensity (RDI). Comparisons between older and younger patients regarding hematologic toxicity and reductions in RDI were based on univariate and multivariate logistic regression analyses.

RESULTS:

The neutropenic complication rate in older patients (45.1%) was not significantly different from that in the younger patients (43.7%). There were also no significant differences in rates of anemia or thrombocytopenia. Approximately 34.0% of the older patients received RDI <85% compared with 20.0% among younger patients (P < .01). Fewer older patients received anthracycline-based chemotherapy (64.3% compared with 83.8% in younger patients, P < .01). Fewer older patients received prophylactic white blood cell colony-stimulating factor (18.4%) compared with younger patients (28.0%) (P < .01).

Conclusions:

There were no significant differences noted with regard to chemotherapy-related hematologic toxicities between older and younger breast cancer patients in this large prospective observational study. This may be explained, in part, by more frequent reductions in RDI and less frequent utilization of anthracyclines among older patients. Cancer 2009. © 2009 American Cancer Society.

Breast cancer incidence in the United States rises significantly with increasing age. Approximately half of the women diagnosed with the disease are aged ≥65 years.1 Benefits from chemotherapy in terms of disease-free and overall survival for women aged 50 to 69 years are substantial, although not as great as those seen in women aged <50 years.2 The benefits from chemotherapy for women aged >65 years have been confirmed, albeit at a price of higher treatment-related mortality compared with younger patients.3 Chemotherapy-associated toxicities have also been shown to increase with age.4 One approach to decreasing chemotherapy-related toxicities has been to reduce the relative dose intensity (RDI) of chemotherapy delivered to patients. With evidence that a threshold of chemotherapy RDI exists below which clinical benefit is substantially undermined,5-8 dose reductions, particularly in the curative setting, should be viewed as a strategy of last resort. Nevertheless, this practice persists,9, 10 particularly with older patients.11 Concerns regarding higher risk to benefit ratios for chemotherapy use in older patients have, in part, fueled the comparative underrepresentation of older patients in clinical trials.12 Likewise, in general practice, despite the National Institutes of Health consensus recommendations,13 chemotherapy use for the management of breast cancer has been shown to decrease with increasing age.14-16 Accordingly, some uncertainty regarding the value of chemotherapy for older women with high-risk breast cancer persists. Determination of the value of chemotherapy is based, in part, on quantifying toxicity rates and extent as well as benefits in a representative population. Understanding the value of chemotherapy for this age group may be compromised by selection bias inherent in clinical trials, which comprise the major source of data on this subject.2, 3, 17 Those deemed eligible to participate in clinical trials may not necessarily be representative of the population of older patients at large, who in general tend to have more comorbidities and take more medications than their younger counterparts. More recent studies assessing trends in chemotherapy use for older breast cancer patients in nonclinical trial registries have not included dose intensity and its potential impact on toxicity in the analysis.14-17 This study was performed to determine the interaction of dose intensity and hematologic toxicity prospectively in older nonrandomized breast cancer patients in a community oncology practice setting.

MATERIALS AND METHODS

Patient Characteristics

A prospective observational study of 115 randomly selected US community oncology practices was undertaken containing a total of 4458 cancer patients starting a new chemotherapy regimen enrolled between March 2002 and October 2005. The study included 1224 consecutive eligible breast cancer patients with nonmetastatic disease. Sites were selected for participation in this patient registry through stratified random sampling incorporating practice size and geographic location. All practice sites were institution review board-approved, and all patients signed informed consent. Anonymous numeric codes were used to ensure patient confidentiality. Patients were enrolled before initiation of at least 4 cycles of chemotherapy. Data were gathered on each patient, including pretreatment patient characteristics such as age, number of comorbidities, and Eastern Cooperative Oncology Group performance status (ECOG PS). Comorbidity assessments on each patient were made using the Charlson Comorbidity Index.18 For the analysis, only the unweighted number of comorbidities, as in 0, 1, or ≥2, and not disease severity was determined. Body surface area was calculated using the Mosteller formula19 incorporating the patient's actual height and weight. Data pertaining to treatment included type of chemotherapy regimen, treatment schedule, and both planned and actual doses for each drug.

Clinical Outcomes

Primary outcome measures included incidence of grade 2 or higher anemia (hemoglobin <10 g/dL), thrombocytopenia (platelets <75 × 1000/mm³), severe neutropenia (SN; defined as neutrophils <500 cells/mm³ or white blood cell count <1000/mm³), febrile neutropenia (FN; neutrophils <1000 cells/mm³ and documented fever/infection), and both planned and actual RDI compared with standard chemotherapy regimens. The definition of standard dose intensity for each drug was the established dose in milligrams per square meter per unit time (week). A comprehensive literature review of articles and reference manuals on chemotherapy was performed to identify the standard or established dose intensity.20-28 Oncology experts were consulted when >1 possible chemotherapy schedule or dose was identified, to provide an opinion on which of the possible schedules or doses should be considered “standard.” To avoid overestimation of standards in the event that >1 potential standard regimen was identified, the more conservative schedule or dosing pattern was selected. The definition of RDI was accounted for by the ratio of either the actual or the planned dose intensity relative to the standard dose intensity. Reductions in RDI, both planned and unplanned, were calculated for each chemotherapeutic agent and averaged for each regimen. RDI for each chemotherapy regimen reflected the average RDI for every chemotherapeutic and myelosuppressive agent contained therein. Planned RDI was defined as that calculated to commence at initiation of treatment as determined by the treating oncologist. The secondary outcome measure in this study included the frequency of actual or planned RDI <85% of the chemotherapy reference standard.

Hematopoietic Growth Factors

Frequency and timing of white blood cell growth factor use was monitored during each chemotherapy cycle. Granulocyte colony-stimulating factor was administered to 99.9% of patients who received growth factors. Only 1 patient received granulocyte macrophage colony-stimulating factor. Administration of colony-stimulating factor (CSF) planned at the beginning of the first cycle or before a neutropenic event within the first cycle constituted primary prophylaxis. All other CSF use was considered reactive. This analysis did not include information on CSF dose or duration of use.

With respect to use of erythropoietin, only frequency of usage was monitored. Frequency referred to the number of patients who received red cell growth factors. Each patient was counted only once. No information regarding the use of platelet-stimulating growth factors was included in this dataset.

Statistical Methods

Within this cohort, patients aged ≥65 years were compared with patients aged <65 years with respect to types of chemotherapy regimens, mean and median RDI of chemotherapy, planned and unplanned reductions in chemotherapy, and variations in type and extent of hematologic toxicities. The comparison of difference between younger and older patients was planned at the outset of the study, with ongoing monitoring of the proportion of older patients enrolled.

Univariate and multivariate logistic regression analyses were used to assess the relation of demographic and clinical variables with primary and secondary outcomes within the older cohort (age ≥65 years). Group comparisons of categorical variables were made using the chi-square test. Cochran-Armitage Trend Test was used for ordinal variables. All covariates were assessed for missing values, and few values were absent (1 value for comorbidities, 2 for ECOG PS, and 10 for stage). For the analysis, unknown values were grouped with either the largest or the most medically relevant category.

Observations with missing outcome values were excluded from the analysis with the respective outcome. For both univariate and multivariate analyses, the RDI categorization was based on values <85% or ≥85%. P < .05 was considered to be statistically significant, and all reported P values are 2-sided.

RESULTS

Of the 1224 patients with stage I to stage III breast cancer, 207 patients were ≥65 years old (grading determined according to the American Joint Committee on Cancer staging system). The mean and median ages for the older patient group were 71.3 years and 70 years, respectively (range, 65-85 years). Table 1 shows a comparison of patient characteristics between the group of patients aged ≥65 years and those patients aged <65 years. Characteristics of the practice sites in which patients were treated did not differ significantly between older and younger breast cancer patients, including geographic region, practice location, and practice size.

Table 1. Comparison of Early Stage Breast Cancer Patients by Age

Category	Age <65 Years (n = 1017), No. (%)	Age ≥65 Years (n = 207), No. (%)	P
ECOG PS indicates Eastern Cooperative Oncology Group performance status; BSA, body surface area; MI, myocardial infarction; CHF, congestive heart failure; RDI, relative dose intensity; CSF, colony-stimulating factor. * Sample size was smaller due to the missing values. † Hematologic toxicities were available for 1203 patients.
Ethnicity			.03
White	799 (78.6)	178 (86.0)
Black	142 (14.0)	15 (7.2)
Other	76 (7.5)	14 (6.8)
Geographic region			.08
Northeast	186 (18.3)	29 (14)
Central	252 (24.8)	65 (31.4)
West Coast	176 (17.3)	27 (13.0)
South	403 (39.6)	86 (41.5)
Practice location			.12
Metropolitan	764 (75.1)	142 (68.6)
Nonmetropolitan	205 (20.2)	55 (26.6)
Unknown	48 (4.7)	10 (4.8)
Practice size			.11
Small	242 (23.8)	48 (23.2)
Small to medium	188 (18.5)	53 (25.6)
Medium	240 (23.6)	50 (24.2)
Medium to large	173 (17.0)	25 (12.1)
Large	174 (17.1)	31 (15.0)
Stage			<.01
I	270 (26.6)	24 (11.6)
II	552 (54.3)	134 (64.7)
III	189 (18.6)	49 (23.7)
Early unspecified	6 (0.6)	0 (0.0)
ECOG PS			<.01
0	817 (80.3)	137 (66.2)
1	184 (18.1)	61 (29.5)
2-4	16 (1.6)	9 (4.3)
Positive estrogen or progesterone receptor			.98
Yes	649 (63.8)	131 (63.3)
No	327 (32.2)	68 (32.9)
Unknown	41 (4.0)	8 (3.9)
Baseline BSA >2 m^2	237 (23.3)	39 (18.8)	.16
History of anemia	75 (7.4)	12 (5.8)	.42
History of MI or CHF*	14 (1.4)	10 (4.9)	<.01
No. of comorbidities			<.01
0	896 (88.1)	151 (72.9)
1	108 (10.6)	46 (22.2)
2-4	13 (1.3)	10 (4.8)
Creatinine >1.5 mg/dL*	5 (0.5)	8 (3.9)	<.01
Bilirubin >1mg/dL*	19 (1.9)	6 (3.0)	.35
History of chemotherapy	79 (7.8)	22 (10.6)	.17
No. of prior treatments			.25
0	938 (92.2)	185 (89.4)
1	51 (5.0)	12 (5.8)
≥2	28 (2.8)	10 (4.8)
Regimen with anthracyclines	852 (83.8)	133 (64.3)	<.01
Planned RDI			.02
≥85%	873 (85.8)	163 (78.7)
<85%	125 (12.3)	36 (17.4)
Unknown	19 (1.9)	8 (3.9)
No. of myelosuppressive drugs ≥2	950 (93.4)	187 (90.3)	.12
Supporting treatment
Prophylactic CSF	285 (28)	38 (18.4)	<.01
CSF during study	635 (62.4)	119 (57.5)	.18
Erythropoietin during study	409 (40.2)	82 (39.6)	.87
Hematologic toxicity during study†
Febrile or severe neutropenia	437 (43.7)	92 (45.1)	.72
Anemia	249 (24.9)	56 (27.5)	.45
Thrombocytopenia	79 (7.9)	22 (10.8)	.18

Data on hematologic toxicities were available for 1203 patients, 204 of whom were aged ≥65 years. There was no statistically significant difference in hematologic toxicity between older and younger patients during the course of this study. Rates of FN or SN were similar between the 2 groups, as were rates of anemia and thrombocytopenia, regardless of toxicity grade. Statistically significant differences were observed between older and younger patients in terms of several clinical characteristics, including disease stage, ECOG PS, number of comorbidities, serum creatinine >1.5 mg/dL, and ethnicity. ECOG PS <2 was observed in 95.7% of the older patients and 98.4% of the younger patients (P = .01). More of the younger patients had an ECOG PS of 0 as compared with the older patients (P < .01). Not surprisingly, the younger patients had fewer comorbidities than the older patients (P < .01), although <5% of the older patients had 2 or more comorbidities. The older patients had a significantly higher proportion of prior cardiac events, including myocardial infarction or congestive heart failure (4.9%) compared with the younger patients (1.4%) (P < .01). In terms of ethnicity, there were fewer non-Caucasian older patients as compared with the younger patients (P < .01). Elevations in serum creatinine were more common among elderly patients, which may compromise renal excretion of several chemotherapeutic agents, increasing the potential for drug toxicity, including myelosuppression.

Treatment variation that differed significantly between older and younger patients included use of anthracycline-containing regimens, planned RDI for chemotherapy, and prophylactic CSF use. There was no significant difference in overall CSF use between older (57.5%) and younger patients (62.4%); however, significantly fewer older patients received prophylactic CSF during the course of the study (18.4%) compared with the younger group of patients (28.0%) (P < .01). There was no difference in the use of exogenous erythropoietin during this study between the older and younger patients (approximately 40% in each group).

Prior chemotherapy use was documented in 10.6% of the older patients and 7.8% of the younger patients (P = .17). Substantial variability in selection of chemotherapy regimen was observed between patients aged ≥65 years compared with patients aged <65 years. Significantly fewer older patients received anthracycline-containing regimens compared with the younger patients (P < .01) (Fig. 1). The most common non–anthracycline-containing regimen used in the older cohort of patients was cyclophosphamide, methotrexate, and fluorouracil at 22.2%. This regimen was used in 7.8% of the younger patients. Overall cyclophosphamide, methotrexate, and fluorouracil use was 10.2%.

Figure 1. Distribution of treatment regimen by age is shown. A indicates doxorubicin; C, cyclophosphamide; E, epirubicin; F, 5-flourouracil; T, docetaxel; M, methotrexate. *The most frequent other (uncommon) adjuvant chemotherapy regimens were single-agent doxorubicin, carboplatin and docetaxel, and carboplatin and paclitaxel.

In the group of older patients, planned RDI was calculated in 199 (96.1%) patients, and actual RDI was calculated in 197 (95.2%). Both planned and actual RDI differed significantly with increasing age (P < .01). The mean actual RDI was 85.6% overall (standard deviation, 17.6%). Just over ⅓ of the older patients (34.0%) with potentially curable breast cancer received an actual RDI <85%. Planned RDI <85% was more common in older patients (17.4%) as compared with younger patients (12.3%; P = .02). Figure 2 shows differences in actual RDI and the planned and unplanned dose intensity reductions of chemotherapy between the age groups and with respect to disease stage. The average actual RDI was significantly higher in younger patients compared with older patients (P < .01).

Figure 2. Averaged relative dose intensity (RDI) by age and stage is shown. Actual RDI represents the mean actual dose intensity administered during the course of treatment compared with standard-dose intensity. Unplanned reduction represents the mean difference between the planned and actual RDI. There was a significant difference noted with regard to the mean planned and actual RDI between the age groups (P < .01). The RDI did not appear to vary significantly between stage of disease groups.

The minimum white blood cell count during the study had a mean of 2800/mm³, with a median of 2100/mm³ and a standard deviation of 2500/mm³. The minimum absolute neutrophil count during the study had a mean of 1400/mm³, with a median of 700/mm³ and a standard deviation of 2000/mm³. Table 2 shows the dose intensity and FN or SN rates among the older patients along with actual RDI <85%. FN or SN over the first 4 cycles of chemotherapy occurred in 92 (45.1%) patients. Rates of FN and SN did not increase with increasing chronological age. Neither did rates of anemia and thrombocytopenia increase with increasing age (Table 3).

Table 2. Dose Intensity and Febrile and/or Severe Neutropenia Among Elderly Breast Cancer Patients*

Category	FN or SN			Actual RDI <85%
	Sample	No. of Events (%)	P	Sample	No. of Events (%)	P
FN indicates febrile neutropenia; SN, severe neutropenia; RDI, relative dose intensity; ECOG PS, Eastern Cooperative Oncology Group performance status; BSA, body surface area; CSF, colony-stimulating factor. * In multivariate analysis after adjusting for baseline absolute neutrophil count, age, prior chemotherapy, and planned RDI anthracycline-containing regimens were found to be associated with more FN or SN (odds ratio [OR], 12.76; 95% confidence interval [95 % CI], 5.61-29.03 [P < .0001]), and prophylactic CSF was found to be associated with less FN or SN (OR, 0.09; 95% CI, 0.03-0.24 [P < .0001]). † Patients with unknown RDI are not included.
All age ≥65 y	204	92 (45.1)	—	197	67 (34.0)
Age, y			.09			.01
65-69	92	44 (47.8)		90	23 (25.6)
70-74	61	30 (49.2)		60	22 (36.7)
75-79	38	16 (42.1)		35	15 (42.9)
80-88	13	2 (15.4)		12	7 (58.3)
ECOG PS			.69			.06
0	134	61 (45.5)		130	38 (29.2)
1	61	28 (45.9)		58	25 (43.1)
2-4	9	3 (33.3)		9	4 (44.4)
History of chemotherapy	21	8 (38.1)	.50	18	5 (27.8)	.56
History of anemia	12	4 (33.3)	.40	11	7 (63.6)	.03
No. of comorbidities			.23			.20
0	148	71 (48.0)		142	45 (31.7)
1	46	17 (37.0)		45	17 (37.8)
2-4	10	4 (40.0)		10	5 (50.0)
Baseline BSA >2 m^2	38	13 (34.2)	.13	37	13 (35.1)	.87
Anthracycline-containing regimen	132	80 (60.6)	<.01	129	43 (33.3)	.78
Planned RDI†			.03			<.01
≥85%	161	78 (48.4)		162	36 (22.2)
<85%	35	10 (28.6)		35	31 (88.6)
Prophylactic CSF	38	6 (15.8)	<.01	38	11 (28.9)	.46

Table 3. Anemia (Hemoglobin <10 g/dL) and Thrombocytopenia (Platelets <75 × 10⁹/L) Among Elderly Patients With Breast Cancer*

Category	Sample	Anemia		Thrombocytopenia
		No. of Events (%)	P	No. of Events (%)	P
ECOG PS indicates Eastern Cooperative Oncology Group performance status; RDI, relative dose intensity. * In multivariate analysis after adjusting for baseline hemoglobin, age, prior chemotherapy, planned RDI, ECOG PS (odds ratio [OR], 1.93; 95% confidence interval [95% CI], 1.04-3.59 [P = .03]), and anthracycline-containing regimens (OR, 3.89; 95% CI, 1.56-9.69 [P < .01]) were found to be associated with more anemia. In multivariate analysis after adjusting for baseline platelets, age, planned RDI, and prior chemotherapy anthracycline-containing regimens (OR, 10.14; 95% CI, 2.02-50.89 [P < .01]) were found to be associated with more thrombocytopenia. † Patients with unknown RDI are not included.
All age ≥65 y	204	56 (27.5)		22 (10.8)
Age, y			.82		.47
65-69	92	23 (25.0)		7 (7.6)
70-74	61	19 (31.1)		8 (13.1)
75-79	38	14 (36.8)		7 (18.4)
80-88	13	0 (0.0)		0 (0.0)
ECOG PS			.02		.55
0	134	29 (21.6)		15 (11.2)
1	61	24 (39.3)		7 (11.5)
2-4	9	3 (33.3)		0 (0.0)
History of chemotherapy	21	4 (19.0)	.36	4 (19.0)	.20
History of anemia	12	5 (41.7)	.26	1 (8.3)	.78
Presence of comorbidities	56	22 (39.3)	.02	7 (12.5)	.63
Anthracycline-containing regimen	132	45 (34.1)	<.01	20 (15.2)	.01
Planned RDI†			.45		.29
≥85%	161	47 (29.2)		19 (11.8)
<85%	35	8 (22.9)		2 (5.7)

Patients who received anthracycline-based regimens were more likely than those who did not to experience FN or SN (P < .01), thrombocytopenia (P < .01), or anemia (P < .01). A significantly greater proportion of older patients who did receive an anthracycline-containing regimen developed neutropenic complications (60.6%) compared with younger anthracycline-treated patients (49.3%) (P = .02). The rates of FN or SN in patients receiving anthracyclines in age ranges 18 to 49 years (49.3%), 50 to 64 years (56.7%), 65 to 69 years (63.6%), 70 to 74 years (66.7%), and 75 to 88 years (66.7%) demonstrated a statistically significant increasing trend (P = .01). Prophylactic use of CSF was associated with significantly less FN or SN (P < .01). A history of ≥1 comorbidities and higher ECOG PS were associated with the development of anemia (P = .02), but not thrombocytopenia, FN, or SN. Prior history of anemia was not significantly associated with the development of treatment-related anemia.

DISCUSSION

In this prospective community oncology practice registry study of 1224 patients with nonmetastatic breast cancer, of whom 207 patients were aged ≥65 years at the time of study entry, there were no statistically significant differences in chemotherapy-related hematologic toxicity between older and younger patients. Specifically, no significant differences were observed between age groups in terms of rates of FN, SN, anemia, or thrombocytopenia.

Several studies have addressed the issue of chemotherapy-related toxicity in older patients with breast cancer.3, 4, 14, 15, 17 In contrast to our results, Muss et al. examined toxicity of adjuvant chemotherapy in older and younger breast cancer patients in the setting of the Cancer and Leukemia Group B experience and reported that patients aged ≥65 years were 68% more likely to develop grade 4 leukopenia as compared with patients aged ≤50 years without increased mortality.17 Chemotherapy-related hematologic toxicities in the older patients in our study may have been abrogated by a significantly greater frequency of reductions in planned and actual RDI of chemotherapy for the older patients. In addition, the decreased use of anthracyclines in this group of patients as compared with the younger patient cohort may have led to fewer chemotherapy-related hematologic toxicities. Differences in treatment approach between older and younger patients in our study may also explain the discrepant conclusions between our study and those of Cancer and Leukemia Group B, where a higher threshold for implementing chemotherapy dose reduction or delay was established through stringent protocol guidelines, with specific predefined indications for such interventions. Our study reflects the patterns of community practice, where often clinical judgment alone regulates maintenance of chemotherapy RDI. Although it is encouraging to find that rates of hematologic toxicities in this study were no higher for older patients, it is concerning that this finding was associated with significant reductions in both planned and actual RDI for chemotherapy. This is all the more notable because the older patients in this study had comparably higher stage disease, wherein there should have been a greater impetus to preserve chemotherapy RDI in the curative setting.

Whereas all patients, regardless of age, in the Cancer and Leukemia Group B study received an anthracycline-containing regimen,17 80.5% of the patients in this study were treated with an anthracycline. Among patients aged ≥65 years, only 64.3% received such treatment, yet the rate of neutropenic complications was significantly higher than that observed in younger anthracycline-treated patients. The FN or SN trend in the setting of anthracycline use rose consistently with increasing age (P = .01), supporting the strong effect of age on anthracycline-related hematologic toxicity.

The older patients in the current study had a significantly greater number of comorbidities, as well as higher overall ECOG PS as compared with the younger patients; however, older breast cancer patients in this study received less prophylactic CSF than their younger counterparts. Therefore, greater supportive care in the older subgroup of patients, at least with respect to prophylactic CSF, could not be said to account for the similarities in observed rates of hematologic toxicity. Erythropoietin was used by approximately 40% of patients regardless of age. Thus, the similar rates of anemia in older and younger patients cannot be ascribed to increased use of erythropoietin in the older patients.

Despite the significant rate of chemotherapy dose reductions in the older cohort used in this study (34%), presumably as a means to reduce risk of treatment-related hematologic and overall toxicities, approximately 40% of patients aged ≥65 years did experience neutropenic complications. Although this rate is similar to that seen in the younger cohort, these toxicities in older patients have been associated with prolonged hospitalization29 and ultimately increased mortality.30

Several limitations of this study must be addressed. Chemotherapy selection was at the discretion of the treating oncologist, with fewer older patients offered anthracyclines as part of a chemotherapy regimen compared with younger patients. This is not surprising, because the risk of cardiotoxicity from anthracyclines approximates 1%,31 and a greater proportion of older patients may be predisposed to development of cardiac complications.32 Indeed, a significantly greater proportion of older patients in this study had a prior history of myocardial infarction or congestive heart failure as compared with their younger counterparts. Before publication of data on the combination of docetaxel and cyclophosphamide,33 clinicians concerned about the potential for chemotherapy-related cardiotoxicity often elected to avoid anthracycline use and use the cyclophosphamide, methotrexate, and fluorouracil regimen. As such, the relatively high representation of regimens such as cyclophosphamide, methotrexate, and fluorouracil observed in this analysis may not necessarily reflect the most current trends in choice of chemotherapy regimens for older patients with breast cancer. Currently, the docetaxel and cyclophosphamide combination represents a viable means of avoiding an anthracycline-containing regimen for management of early stage breast cancer. A follow-up report on the docetaxel and cyclophosphamide combination34 supported the tolerability of this regimen in older patients with breast cancer.

The older patients in this study were not as ethnically diverse as the younger patients, and this may call into question the generalizability of our conclusions. Surveillance, Epidemiology, and End Results data demonstrate that whereas age-adjusted incidence for breast cancer is higher in whites than blacks, age-adjusted breast cancer mortality rates are higher for blacks, regardless of age, as compared with whites.35 Furthermore, overall mortality rates secondary to comorbidities are higher in blacks compared with whites.36, 37 Regardless of age, American Indian/Alaska natives and Asians have some of the lowest rates of breast cancer incidence.38 Therefore, the proportional distribution of older breast cancer patients with respect to race in this study is representative of national trends, and as such our findings should indeed be generalizable.

Increasing age has been identified as an independent risk factor for nonreceipt of standard treatment for breast cancer.39 This practice has been observed despite the survival advantage conferred in this population by chemotherapy, particularly for patients with estrogen receptor-negative disease.15 Although this study highlights the significant dose reductions in chemotherapy RDI experienced by older breast cancer patients, it does not address the more widespread issue of nonreceipt of chemotherapy in this patient population. Comparison of this study with the Cancer and Leukemia Group B13 data suggests that a greater proportion of older breast cancer patients, when they do receive chemotherapy, do so outside of a clinical trial setting. Because the plan from the outset of this large registry study was to compare data on older and younger patients with respect to chemotherapy-related dosing and toxicity, the proportion of patients aged ≥65 years old was monitored throughout the enrollment period. Older patients represented 33% to 40% of the entire study population across various tumor types. In the breast cancer cohort the proportion of older patients with nonmetastatic disease was 16.9%. This is substantially greater than the 10% of patients aged ≥65 years represented in the Cancer and Leukemia Group B experience.17 Such differences in the proportions of older patients per study reflect the underrepresentation of older patients in clinical trials. Yet despite this study having a more than 67% greater representation of older patients than the Cancer and Leukemia Group B study, the mean and median age of the patients aged ≥65 years in this study were 71.3 years and 70 years (range, 65-85 years), respectively. This study is, therefore, reflective of the “young old” population, in which a lower prevalence of age-related events is expected. As such, this study is highly selective, not only with respect to functional status and comorbidity but also in terms of age. The ongoing participation in clinical trials of older patients with various functional levels is needed to provide quality evidence on risks and benefits in this highly diverse group of patients.

Conflict of Interest Disclosures

Dr. Dale acts as a consultant and receives research support from Amgen, which had a financial interest in the topic of this article.

Dr. Crawford acts as a consultant for Amgen, Ortho, Genentech, and Chugai and takes part in Educational Programs for Amgen, Genentech, and EMD Serono.

Dr. Lyman is the principal investigator for a research grant awarded from Amgen to Duke University. He is also a member of the Speakers' Bureau for Amgen and Genomic Health.