Adjuvant Chemotherapy in Older and Younger Women With Lymph Node–Positive Breast Cancer FREE

The incidence of breast cancer increases with increasing age, and almost half of all new breast cancers in the United States now occur in women 65 years of age or older.¹ Systemic adjuvant chemotherapy in women with early-stage breast cancer significantly improves both relapse-free and overall survival for women aged 50 to 69 years old, but data are lacking for women aged 70 years or older.² Nevertheless, available data suggest that systemic adjuvant chemotherapy may be significantly underused in older patients,³ even though for many patients in this setting, chemotherapy may improve survival. Moreover, even when chemotherapy is administered to older patients, inappropriate dose reductions are frequently made that may decrease effectiveness.⁴

The potential for increased chemotherapy-related toxic effects in older patients leads to important concerns. For instance, renal function and marrow reserve decrease with age and can increase the risk of toxic effects from treatment regimens that include myelosuppressive agents or agents such as methotrexate that are renally excreted.⁵ Dose modification using creatinine clearance values can minimize toxic effects.⁶ Also, data from small trials and retrospective analyses of larger trials suggest that older women in good health tolerate chemotherapy, including anthracycline-based regimens, with toxic effects profiles similar to those of younger patients.^7- 10

Recently, randomized clinical trials of combination chemotherapy regimens in the adjuvant setting have shown that dose-intensive chemotherapy regimens and regimens incorporating newer antineoplastics, such as taxanes, are associated with significant improvements in both relapse-free and overall survival when compared with older, more established treatments. These more effective regimens have often included anthracyclines and include trials comparing anthracyclines with nonanthracycline regimens,¹¹ larger vs smaller doses of anthracyclines,¹² anthracyclines with vs without the inclusion of taxanes,¹³ and dose-dense regimens that give the same doses of chemotherapy in a shorter time.¹⁴ The benefits and risks of these more toxic regimens have not been adequately explored in older patients because older patients have been underrepresented in clinical trials.^15- 16 To determine how older patients fared with more aggressive systemic adjuvant chemotherapy regimens, the Cancer and Leukemia Group B (CALGB) retrospectively reviewed 4 randomized clinical trials of treatments for lymph node–positive breast cancer that accrued patients from 1975 to 1999. These trials compared more aggressive with less aggressive chemotherapy regimens.

Data were obtained from 4 CALGB randomized trials designed for patients with node-positive breast cancer. All trials compared at least 2 chemotherapy regimens that differed by dose level, dose intensity, or regimen (Table 1). All trials required informed consent based on federal, state, and institutional guidelines. The CALGB 7581 trial randomly assigned patients to receive cyclophosphamide, methotrexate, and fluorouracil (CMF), CMF plus vincristine and prednisone (CMF+VP), or CMF plus the methanol extraction residue of bacillus Calmette-Guérin (CMF+MER). Details concerning this trial have been previously published.¹⁷ In a multivariate proportional hazards model, CMF+VP was significantly superior to the other 2 arms combined (CMF and CMF+MER) in improving disease-free survival (P = .009); overall survival was not significantly different among the groups. Endocrine therapy was not addressed in this trial and it is unlikely that any of these patients received adjuvant tamoxifen.

The CALGB 8082 trial was a randomized trial that compared CMF+VP or CMF+VP followed by a doxorubicin-based regimen: vinblastine, doxorubicin, thiotepa, and fluoxymesterone (Halotestin) (VATH). All patients were given a 6-week induction course of CMF+VP and then were randomly assigned to receive 1 of 2 CMF+VP consolidation regimens. One consisted of 6 months of conventional 2-week blocks of CMF+VP therapy separated by 2-week rest periods, and the second consisted of 2 additional 6-week blocks separated by 6-week rest periods. At the end of the first 8 months of CMF+VP therapy, patients were again randomly assigned to continue CMF+VP for 6 more months or to receive 6 cycles of escalating doses of VATH. The results of this trial have been previously published.¹⁸ There was no statistical difference in disease-free survival between the 2 CMF+VP consolidations regimens. Compared with continuing CMF+VP, VATH intensification was associated with significantly improved disease-free survival (P = .004) and overall survival (P = .04). Endocrine therapy was not addressed in this trial but it is unlikely that any of these patients received adjuvant tamoxifen.

In the CALGB 8541 trial, patients were randomly assigned to receive 1 of 3 regimens of cyclophosphamide, doxorubicin, and fluorouracil (CAF) that differed in dose duration and dose intensity. The high-dose group received cyclophosphamide at 600 mg/m², doxorubicin at 60 mg/m², and fluorouracil at 600 mg/m², every 4 weeks for 4 cycles, with fluorouracil repeated on day 8 of each cycle. The moderate-dose group received 400 mg/m² of cyclophosphamide, 40 mg/m² of doxorubicin, and 400 mg/m² of fluorouracil in the same schedule as the high-dose group but for 6 cycles. The low-dose group received exactly half the doses of the high-dose group, on the same schedule. Details of this trial have been previously published.¹⁹ Patients treated with moderate- or higher-dose intensity of CAF had significantly longer disease-free survival (P<.001) and overall survival (P = .004) compared with those treated with lower doses. There were no differences in disease-free survival or overall survival between the moderate- and high-dose groups, but a subsequent analysis showed that patients whose tumors were human epidermal growth factor receptor 2 (ERBB2; formerly HER-2 or HER-2/neu)–positive had significantly improved disease-free survival and overall survival when treated with the high-dose regimen compared with the moderate- and lower-dose regimens.²⁰ For patients who were ERBB2-negative, there were no differences in disease-free and overall survival among the 3 different treatment arms. In 1998, an addendum to the protocol required the addition of tamoxifen after CAF in patients who were perimenopausal or postmenopausal and who had estrogen receptor–positive tumors. Patients already participating in this study who met these criteria could be given tamoxifen at a time decided by their treating physician. All subsequent accruals, however, were required to begin tamoxifen therapy. Tamoxifen was given to 18% of patients aged 50 years or younger, 47% of patients aged 51 to 64 years, and 50% of patients aged 65 years or older.

The CALGB 9344 trial (Intergroup Trial 0148) used a 3 × 2 factorial design in which patients were randomly assigned first to 1 of 3 dose levels of doxorubicin (60, 75, or 90 mg/m²) and a fixed dose of cyclophosphamide (600 mg/m²). Both agents were administered every 3 weeks for 4 cycles (AC). Subsequently, all patients were randomly assigned to either receive or not receive paclitaxel (175 mg/m²), given every 3 weeks for 4 cycles. Patients with estrogen receptor–positive tumors in this trial were required to take tamoxifen, 20 mg/d for 5 years. Details of this trial have been previously published.¹³ There were no differences in outcome among patients randomized to the 3 different dose levels of doxorubicin. Patients randomized to paclitaxel, however, had significantly improved 5-year disease-free survival (P = .002) and 5-year overall survival (P = .006).

A total of 6593 patients were accrued to these trials. Of these, 6489 received protocol therapy; data for 2 patients were missing date of birth and therefore are not evaluable for age. The resulting sample of 6487 patients is the basis of the results reported herein.

Disease-free survival was defined as the time from study entry until first recurrence of breast cancer or death due to breast cancer; therefore, a treatment failure refers to locoregional or distant recurrence or death from breast cancer. Patients who died of causes other than breast cancer and without recurrence were censored at their date of death, and patients still alive without recurrence were censored at the last date that they were known to be disease-free. Overall survival was defined as the time from study entry until death due to any cause. Surviving patients were censored at the last date they were known to be alive. Age refers to patient age at study enrollment.

Proportional hazards regressions with Wald χ² tests were used to multivariately model and assess the relationship between disease-free survival (or overall survival) and treatment (more or less chemotherapy), standard clinical variables (ie, age at study entry [≤50, 51-64, or ≥65 years], number of positive nodes, tumor size, estrogen receptor status [negative or positive], tamoxifen use [yes or no], and treatment study [CALGB 7581, 8082, 8541, or 9344]). We also assessed whether there was any differential benefit of chemotherapy by age (a chemotherapy-age interaction term). Time-to-event distributions were calculated using the Kaplan-Meier product-limit method. Proportions were compared using contingency table analysis; their 95% confidence intervals (CIs) used exact binomial methods. All P values are 2-sided and P<.05 was considered statistically significant. Analyses were carried out using SAS statistical software, version 8.02 (SAS Institute Inc, Cary, NC).

Trial characteristics and age distribution of the study sample are presented in Table 1. The 4 trials spanned a 24-year period and included a total of 6593 patients initially accrued and 6487 patients comprising the study sample. Median follow-up time ranged from 6.0 years (CALGB 9344) to 23.5 years (CALGB 7581), with an overall median of 9.6 years. Patients aged 65 years or older comprised 8% (542 patients) of the study sample, and 2% (159 patients) were at least 70 years old.

Clinical characteristics by age are presented in Table 2. Tumor sizes were similar across the age groups; however, the proportions of patients with 10 or more positive lymph nodes at the time of entry were significantly different. Twenty-five percent of patients aged 65 years or older had 10 or more positive nodes compared with 17% of patients aged 51 to 64 years and 11% of patients aged 50 years or younger. Estrogen receptor status was similar among age groups; progesterone receptor data were missing for 30% of patients, precluding meaningful comparisons among age groups. Tamoxifen data were missing for 28% of patients; at the time that CALGB 7581 and 8082 were accruing patients (1975-1984), tamoxifen was not considered standard therapy for patients with estrogen receptor– or progesterone receptor–positive tumors and no recommendations were made for its use in either of these trials, nor were data pertaining to tamoxifen use collected. In CALGB 8541, 97% of patients had estrogen receptor status recorded and estrogen receptor–positive patients included 61% of those aged 50 years or younger, 66% of those aged 51 to 64 years, and 62% of those aged 65 years or older. Tamoxifen use in these 3 age groups was 18%, 47%, and 50%, respectively, and these data were available for 99% of patients. In CALGB 9344, 98% of patients had estrogen receptor status recorded and estrogen receptor–positive patients included 57% of those aged 50 years or younger, 62% of those aged 51 to 64 years, and 62% of those aged 65 years or older. Tamoxifen use in these 3 age groups was 67%, 72%, and 71%, respectively.

Table 3 shows treatment-related mortality by age. The overall treatment-related mortality was 0.5% (95% CI, 0.4%-0.7%). There was a significant relationship between age and death due to protocol therapy. Older patients had higher chemotherapy-related mortality, and the incidence of treatment-related mortality increased linearly with increasing age: 0.2% (≤50 years), 0.7% (51-64 years), and 1.5% (≥65 years) (P<.001).

Table 4 shows the multivariate proportional hazards regression model, which relates patient age and degree of chemotherapy to disease-free and overall survival after adjustment for standard clinical variables and treatment study. There was no interaction of chemotherapy and age for either disease-free survival or overall survival. The standard prognostic variables were statistically significant for both disease-free and overall survival. More positive nodes, larger tumor size, no tamoxifen use, and having an estrogen receptor–negative tumor were prognosticators of shortened disease-free and overall survival. There was a greater risk of treatment failure and shorter survival for patients entered in studies other than CALGB 9344.

Degree of chemotherapy significantly related to disease-free survival and overall survival (P<.001). The observed reduction in hazard of failure of relapse was 22% for patients who received more chemotherapy vs those who received less chemotherapy (hazard ratio, 0.78; 95% CI, 0.72-0.85). Patient age did not relate to disease-free survival. Figure 1 shows the effects of chemotherapy on disease-free survival by age group. Regardless of age, those who received more chemotherapy consistently had longer disease-free survival. The reduction in hazard of failure attributed to more chemotherapy compared with less chemotherapy was 18% (95% CI, 9%-27%) for women aged 50 years or younger, 20% (95% CI, 8%-30%) for women aged 51 to 64 years, and 42% (95% CI, 22%-56%) for women aged 65 years or older (Table 5).

More chemotherapy, after adjustment for standard clinical variables, was related to significantly longer overall survival (P<.001). More chemotherapy was associated with a hazard of death that was 18% lower than that of less chemotherapy (Table 4). Figure 2 shows the concomitant relationship between degree of chemotherapy and patient age with overall survival. Specifically, regardless of age, those who received more chemotherapy had better overall survival; regardless of degree of chemotherapy, younger age was associated with better overall survival. The reduction in hazard of death due to more chemotherapy compared with less chemotherapy was 17% (95% CI, 6%-27%) for women aged 50 years or younger, 16% (95% CI, 4%-27%) for women aged 51 to 64 years, and 27% (95% CI, 5%-44%) for women aged 65 years or older. Survival and disease status by age is presented in Table 5. Older patients were more likely to have died of either breast cancer– or non–breast cancer–related causes. The reduction in the hazard ratios for patients receiving more compared with less chemotherapy was similar among all age groups.

This study showed that older patients in reasonably good health who met the rigorous eligibility criteria needed for inclusion into these randomized trials derived similar benefits from more chemotherapy treatment as did younger patients. Older patients in these trials were at higher risk for breast cancer recurrence, as evidenced by the higher percentage of older patients with involvement of 10 or more lymph nodes. This finding, as well as the fact that only 8% of patients entered in these trials were aged 65 years or older, underscore what is probably substantial age bias when offering patients clinical trials. Kemeny and colleagues²¹ have shown that age bias remains a significant independent cause of oncologists’ reluctance to offer participation in clinical trials to older patients; only 34% of women aged 65 years or older with stage II breast cancer²² and eligible for a clinical trial in their institution were offered participation, compared with 68% of women younger than 65 years.

Not all older patients with node-positive breast cancer are good candidates for chemotherapy, and such treatment may be inappropriate for older patients with frailty or significant comorbidity. Comorbidity increases with age and can have a profound effect on the survival of patients with breast cancer.^23- 24 Among 1312 patients aged 55 years or older with breast cancer, Yancik et al²³ found 1 or more major comorbidities in approximately 14% of those aged 55 to 59 years, in about 30% of those aged 65 to 69 years, and in about 48% of those aged 75 to 79 years. Nevertheless, it appears almost certain that for many older patients seen during the period that the trials analyzed in this study encompassed, age bias played a major role in whether trial participation was offered. At present, women aged 65, 75, and 85 years in generally good health can expect to live, on average, an additional 20, 12, and 6 years, respectively²⁵ ; if they were to develop high-risk breast cancer, many of these women would likely derive major benefits from adjuvant chemotherapy.

In this study, older patients experienced a higher treatment-related mortality of 1.5% compared with younger patients (0.2%-0.7%). This increase in chemotherapy-associated toxic effects with older age has been noted by others.²⁶ Hospitalizations resulting from chemotherapy toxic effects in 35 060 women with stage I to IV breast cancer aged 65 years or older were analyzed by Du and colleagues.²⁷ Neutropenia, fever, thrombocytopenia, and other adverse effects of treatment resulted in hospitalizations for 6.3% of patients with stage I and 8.1% with stage II breast cancer. The hospitalization rate increased with increasing comorbidity and the use of anthracyclines but was not associated with age. Crivellari and colleagues²⁶ noted that older patients were much more likely to have grade 3 treatment-related toxic effects with a CMF adjuvant therapy regimen, but they noted no grade 4 toxic effects.

The overall risk reduction in breast cancer relapse in this analysis was 22% lower for patients treated with higher-dose regimens. These data reflect the results of the individual trials themselves. Although these trials did not stratify patients by age, the older cohort of patients showed a hazard reduction due to more chemotherapy that was similar to that in younger patients. This study was complicated by the fact that tamoxifen was not used in the 2 early trials but was used in the 2 later trials (CALGB 8541 and 9344) However, even in these 2 later trials, in which data on tamoxifen use were available, the more intensive chemotherapy regimen was still a significant factor in decreasing the risk of relapse in our multivariate analysis. This indicates that there is added value to receiving higher-dose chemotherapy regimens, even in patients receiving tamoxifen. Considerable uncertainty still exists concerning the added value of chemotherapy in older, node-positive, postmenopausal, hormone receptor–positive patients given tamoxifen. Data from the overview analysis, however, show a significant improvement in disease-free and overall survival for patients treated with tamoxifen and chemotherapy compared with those treated with tamoxifen alone.²⁸ The added value of chemotherapy in older women who receive tamoxifen is influenced greatly by comorbidity and life expectancy. Extermann and colleagues²⁹ have developed models for estimating the benefits of chemotherapy in hormone receptor–positive older women and have demonstrated that high risks of recurrence are needed to achieve even small survival benefits for adjuvant chemotherapy. For example, to reduce mortality risk at 10 years by 1% with chemotherapy, the risk of breast recurrence at 10 years had to be at least 25% for a 75-year-old in average health, a recurrence risk that may not be exceeded by all node-positive women. These data suggest that chemotherapy for older women with hormone receptor–positive breast cancer should be offered only to node-positive patients who are in reasonable health, with a high risk of recurrence and a life expectancy of more than 5 years. Results of the current study suggest that higher-dose chemotherapy regimens as used in younger patients are worthy of consideration in the older patient population. Older node-negative patients are unlikely to benefit from chemotherapy unless they have large hormone receptor–positive tumors with adverse pathologic characteristics (lymphovascular invasion or high tumor grade) or hormone receptor–negative tumors larger than 2 cm.³⁰ A model that incorporates age, health status, and tumor characteristics can be helpful in estimating the benefit of adjuvant chemotherapy in older patients (for example, see http://www.adjuvantonline.com). Our data also show a significant survival benefit for more intensive chemotherapy in healthy older patients that met the stringent eligibility criteria for these trials. Older patients in this study more frequently died of non–breast cancer causes than younger patients, but even in the older age group, 73% of deaths were due to breast cancer. Nevertheless, only 2% of patients in this study were aged 70 years or older, and caution should be exercised in extrapolating these data to patients aged 75 years or older, who have shorter life expectancies and more comorbidities than patients aged 66 to 70 years.

Our study adds to the increasing number of trials that suggest that older patients in fair to good health tolerate standard chemotherapy regimens, and even more intensive regimens, almost as well as younger patients.³¹ Moreover, and more importantly, this study suggests that the added value gained from more intensive chemotherapy regimens commonly used in the adjuvant setting might be shared by older patients and not limited to younger age groups. A sobering finding from this analysis is the observation that only 8% of patients entered in the trials analyzed in this study were aged 65 years or older; about 50% of new breast cancer diagnoses occur in women in this older age group. Although good clinical judgment likely played a role in limiting the offering of these trials to many older patients, it is likely that age bias remained a major factor for offering older women clinical trial participation. For example, older patients entered in these 4 CALGB trials had a significantly higher number of positive lymph nodes than younger patients, suggesting that physicians were wary of offering these trials to lower-risk, node-positive older patients.

The majority of current chemotherapy trials for patients with node-positive cancer build on previous gains and include dose-dense regimens, new chemotherapeutic agents or biologics, or schedules of drug administration that many clinicians will perceive as being too toxic for older patients. The data from this study should help to encourage clinicians to offer healthy older patients participation in newer trials, because healthy older patients are likely to derive similar treatment benefits as younger patients. However, depending on the specific research question, older patients need to be carefully counseled about a higher risk of treatment-related toxicity. Trials exploring new approaches to adjuvant chemotherapy for older patients are now in progress. Older patients with high-risk early breast cancer who are in otherwise good health should be offered participation in ongoing clinical trials of adjuvant chemotherapy.