The effect of under-treatment of breast cancer in women 80 years of age and older

1. Background 

The average life expectancy in the United States has been rising nearly continuously since 1900. In 2005, the mean life expectancy of a woman at birth was 80 years [1]. This aging of the population is projected to continue; in the next 25 years the number of American women over the age of 65 years is expected to increase 104% and the number of women 85 years and older is also projected to double [2], [3].

As with most cancers, the risk of developing breast cancer increases with age from 1 in 209 for women younger than 40 to 1 in 14 for women over the age of 70 years old [4]. Despite the increase in the elderly (patients age 65 and older) and very elderly (patients age 80 and older) population, safety, efficacy and necessity of therapeutic modalities in the very elderly breast cancer patients remain under-studied. Retrospective studies have demonstrated that older patients are less likely to receive standard care for breast cancer [5], [6], [7]. Furthermore, the mortality rates for these patients in the last decades have shown less improvement compared to younger patients [8]. This may suggest an adverse effect on overall and disease-free survival in elderly women in whom comprehensive care is compromised: the impact of which, has not been previously evaluated in the very elderly population. As elderly women have historically been excluded from prospective clinical trials, treatment guidelines are usually based on results of studies in a younger population. Therefore, there is an ongoing debate on the optimal treatment for very elderly patients.

An improved understanding of management guidelines may decrease the incidence of premature cancer death or avoidable morbidity in an elderly but otherwise healthy population. Therefore this study was performed to assess the efficacy and necessity of individual therapies in the very elderly breast cancer patient.

2. Methods 

After obtaining Institutional Review Board approval a retrospective analysis was performed on data collected from the medical records of newly diagnosed breast cancer patients 80 years and older evaluated at the MD Anderson Cancer Center between September 1, 1989 and September 1, 2004. Any patient who received breast cancer counseling (and ultimately refused care) or any component of multi-modality breast cancer treatment in our institution was included. Patients presenting with a recurrent breast cancer were excluded from this analysis. Two hundred twelve unique patients were identified. Data extracted from charts included patient demographics, radiological data including the method of diagnosis, clinical and pathologic stage at time of diagnosis, and tumor receptor status. The American Society of Anesthesiologists (ASA) risk assessment was implemented to quantify physical status [9]. Incidences of recurrence or metastasis were documented. Time of death and cause (disease related) were recorded from the chart. The treatment plans in our cohort were compared with accepted recommendations for care of post-menopausal women at the time of diagnosis [7], [10]. The outcomes of patients who received complete therapy based on the accepted recommendations at the time of diagnosis were then compared to the outcomes of women in whom less aggressive therapy was performed.

The associations between the status of standard care treatment, the year of diagnosis, and the patient prognostic factors were analyzed using Fisher's exact test or Chi-square tested. Kaplan–Meier curves were plotted for estimating time to death, time to disease specific death, and time to local and regional recurrence. The median survival times with the 95% confidence intervals and the survival rates at 5 and 10 years with the 95% confidence intervals were provided for each time to event endpoint. The log-rank test was used to compare the difference in survival between the groups. Cox proportional hazard models were utilized for multivariate analysis.

3. Results 

Two hundred twelve patients were identified from our institutional database, patient demographics and treatment are described in Table 1. The median age was 83.5 years (range 80–97). Overall survival in the entire cohort was 7.28 years with a median follow up of 4 years for patients still alive at the end of the study period. Ninety-two (48.4%) patients were diagnosed with stage 1 disease, ten patients presented with stage IV disease. Forty three percent of patients received all treatment indicated for disease stage. The remaining 57% were under-treated according to institutional and national guidelines, 39% did not undergo surgical lymph node evaluation by sentinel lymph node biopsy or axillary lymph node dissection. Eighteen patients were treated with systemic chemotherapy either as a sole treatment (n=14) or in combination with surgical and hormonal treatment (n=2), radiation treatment (n=1) or radiation and hormonal treatment (n=1). These numbers are too small to allow for individual analysis of the impact of chemotherapy on outcome and therefore this treatment modality was left out of further analysis. Five patients (2%) refused all recommended therapy. Ninety-five percent (188/198) of surgical candidates who opted for any treatment underwent surgery. There are no operative deaths to report. Eighty two percent (112/137) of eligible women received hormonal therapy. Patients with fewer comorbidities (ASA I/II) were statistically more likely to undergo complete recommended treatment (P=0.005) than patients with poorer physical status (ASA III/IV).

Table 1. Patient characteristics.

	Variable	Levels	Count	Percent
	Race	White	171	80.7
		Black	29	13.7
		Hispanic	5	2.4
		Asian	7	3.3
	ASA class	I	26	12.6
		II	135	65.5
		III	43	20.9
		IV	2	1.0
	Tumor stage	I	92	48.4
		II	57	30.0
		III	31	16.3
		IV	10	5.3
	ER_PR	ER+ only	35	22.0
		ER+/PR+	97	61.0
		ER−/PR−	21	13.2
		PR+ only	6	3.8
	Standard care	No	120	57.1
		Yes	90	42.9
	Treatment	H/Xrt/HXr	15	7.1
		No Treatment	46	21.7
		Chemotherapy only	14	6.6
		Surgery only	50	23.8
		Surgery+hormonal therapy	50	23.8
		Surgery+radiotherapy	35	16.7
	Surgical lymph node evaluation	SLN	38	18.1
		SLN+axillary dissection	8	3.8
		Axillary dissection	82	39.1
		Not done	82	39.1

ER–PR: estrogen receptor–progesteron receptor, H: hormonal treatment, Xrt: radiotherapy, HXr: hormonal treatment and radiotherapy, SLN: sentinel lymph node biopsy.

When comparing those patients who underwent recommended treatment to those who did not, race, year of diagnosis and tumor stage did not differ between groups (P>0.05). However when analyzing these variables for individual treatment modalities there were differences present (Table 2). The likelihood of receiving no treatment or a unimodality treatment increased with age (Table 3).

Table 2. Distribution of certain variables in different treatment groups.

	Variable	Levels	Not treated (%)	Surgery only (%)	Surg+H (%)	Surg+Xrt (%)	H/Xrt/HXr (%)	P
	Race	White	46 (27.2)	42 (24.9)	43 (25.4)	32 (18.9)	6 (3.6)	0.00045
		Other	14 (34.1)	8 (19.5)	7 (17.1)	3 (7.3)	9 (22.0)
	ASA class	I/II	46 (28.9)	42 (26.4)	36 (22.6)	30 (18.9)	5 (3.1)	0.00097
		III/IV	13 (28.9)	6 (13.3)	12 (26.7)	5 (11.1)	9 (20)
	Tumor stage	III/IV	22 (53.7)	2 (4.9)	3 (7.3)	3 (7.3)	11 (26.8)	<0.0001
		I/II	37 (25)	38 (25.7)	45 (30.4)	24 (16.2)	4 (2.7)
	Year of diagnosis	89–96	17 (23)	17 (23)	24 (32.4)	11 (14.9)	5 (6.8)	0.27701
		97–04	43 (31.6)	33 (24.3)	26 (19.1)	24 (17.6)	10 (7.4)

Surg: surgical treatment, H: hormonal treatment, Xrt: radiotherapy, HXr: hormonal treatment and radiotherapy.

Table 3. Median age in years in different treatment categories (P=0.0243).

	Treatment	N	Median	Lower quartile	Upper quartile
	None	60	83.51	81.64	85.61
	Surgery only	50	82.23	80.68	84.85
	H/Xrt/HXr	15	83.76	82.53	88.54
	Surgery+hormonal treatment	50	82.61	80.91	86.04
	Surgery+radiotherapy	35	81.69	80.81	83.60

H: hormonal treatment, Xrt: radiotherapy, HXR: hormonal+radiotherapy.

Twenty-five patients (12%) died of their disease and twenty-six (12%) patients suffered local regional recurrence. Overall and disease-free survival was not significantly associated with under-treatment of disease (P=0.09 and P=0.94 respectively). However, when stratified by treatments administered, patients who received only one modality of therapy had significantly worse outcomes with respect to local regional control and disease specific survival. Women who underwent hormonal therapy only, demonstrated decreased disease specific survival (P<0.001 respectively) (Fig. 1) compared with patients who received combined therapy. Women who underwent partial mastectomy without radiation treatment experienced a significant increase in local regional recurrence (P=0.045) (Fig. 2). The relatively low number of patients in the individual treatments groups (especially those treated with hormonal and/or radiotherapeutic treatment only) did not allow for a comparison of median survival rates between groups.

View Large Image Download to PowerPoint

Fig. 1. Disease specific survival in patients undergoing hormonal therapy alone versus hormonal therapy with additional treatment modalities (i.e. surgery, or a combination of surgery and radiotherapy).

View Large Image Download to PowerPoint

Fig. 2. Local regional recurrence in patients treated with surgery alone versus surgery and radiation.

Patients who did not undergo surgical lymph node evaluation showed decreased overall and disease specific survival (P<0.001 and P=0.04 respectively) (Fig. 3).

View Large Image Download to PowerPoint

Fig. 3. Disease specific survival in patients undergoing surgical lymph node evaluation (either by sentinel node procedure, formal axillary dissection or both) versus no evaluation.

4. Discussion 

This study shows that the majority of very elderly women are still offered less than standard and unimodality instead of multi-modality treatment, resulting in unnecessary local regional recurrence as well as an increased possibility of risk of disease related death.

Several factors have been named in the literature as a basis for substandard treatment: older patients have a higher prevalence of comorbidities and lower life expectancy compared to younger patients [11], treatment may be perceived to be less effective in the elderly and present a higher risk of adverse events, tumor biology in the elderly may be different from that at a younger age and social isolation may make it more difficult for patients to cope with cancer and the accompanying treatment at a later age [12].

But, since most of these assumptions are not evidence based, the question remains whether treatment should be omitted based on the fear that it may cause unnecessary damage and diminish quality of health.

In the present study one factor associated with less then standard treatment was an advanced ASA class. In this cohort nearly 80% of patients were classified as ASA I or II, indicating that, most of our patients are vital, healthy women. This may be a reflection on the referral patterns at our institution, a significant proportion of elderly women will not be referred to a specialized team. Nevertheless, optimal cancer treatment is warranted even at this seemingly late age as a healthy 80-year-old woman still has a life expectancy of 13 years and even women suffering from severe comorbidities can expect to live another 4 years at this age [13]. In view of the relatively low number of patients with an advanced ASA class in our cohort, it is difficult to comment on the effect of more severe comorbidities on overall and disease-free survival. However, we have already shown that women in this age category are able to tolerate different types of breast cancer treatment without significant increase in morbidity over the general, younger population [14].

There seems to be little debate on the benefits of primary surgery for the treatment of breast cancer in healthy elderly patients [15], [16], [17], [18]. The increased psychosocial morbidity that is present 3 months after surgical treatment has resolved after 2 years and primary endocrine treatment with tamoxifen is associated with inferior local disease control compared to surgery [15]. Our results show that the same holds true for the very elderly population. They also confirm the recent findings showing that adjuvant tamoxifen therapy in older women with early breast cancer increased both overall and cancer specific survival [19].

A subject that remains controversial in everyday clinical practice is whether or not to perform surgical axillary node evaluation in elderly patients. The benefit of sentinel lymph node biopsy and/or axillary node dissection has been extensively shown for younger women with breast cancer in terms of tumor control and disease specific survival [20], [21]. However, in octogenarians it is often omitted because of the perceived extra anesthesiological burden and prolonged duration of surgery. This common practice is supported by the results of studies showing no effect of under-treatment of women over the age of 70 on disease-free survival [22]. In our population axillary node evaluation was omitted in 39% of patients. It is not surprising that these patients show decreased overall survival. The reason for omitting axillary node evaluation is quite often the presence of comorbidities and poor overall health and this in itself will lead to decreased overall survival. But as 80% of our patients were ASA I or II and thus relatively healthy, it is debatable how poor the health of those patients not receiving axillary node evaluation was. Strikingly, our results show that disease specific survival is increased after surgical lymph node evaluation (either by sentinel node procedure, formal axillary dissection or both). This is in accordance with findings by Owusu et al. who showed that receipt of guideline therapy (including axillary lymph node dissection) increased 5-year breast cancer specific survival in women >75 years of age [23]. It stresses the importance of guideline therapy including surgical lymph node evaluation in octogenarians particularly for those who can undergo this therapy.

Some studies have investigated the feasibility and possible benefits of adjuvant systemic chemotherapy in elderly breast cancer patients [24], [25], [26]. But although chemotherapy added to tamoxifen has been found to be more effective than tamoxifen alone in women 50–69 years, data on the efficacy of chemotherapy in patients over the age of 70 are too scarce too determine optimal chemotherapy regimes and doses for patients in this age category [25], [27]. The results of this study cannot be helpful in answering that question for the very elderly since the number of patients receiving chemotherapy plus another modality treatment in this study were simply too low to allow analysis.

The issue of whether or not to add radiotherapy after a surgical procedure seems difficult to resolve. It is often argued that the very elderly patients do not have enough life years left, to experience any benefit from radiotherapy in terms of a reduced local recurrence rate. Valassiadou recently showed that radiotherapy could be successfully omitted after breast conserving surgery in women over the age of 70 years old [28]. Provided the tumors were ER positive and women received adjuvant tamoxifen, the local recurrence rate in these patients was comparable to that of younger patients treated by conventional therapy. However, after reviewing treatment outcome in a much larger series of 8724 women aged 70 years or older, Smith et al. found that the addition of radiotherapy after breast conserving treatment did reduce the risk of a locoregional recurrence and subsequent mastectomy [29]. Patients aged 70–79 years with minimal comorbidity were most likely to benefit and older patients with substantial comorbidity were least likely to benefit. This seems logical as patients with minimal comorbidity have a longer life expectancy and thus have more time to develop a local recurrence when treated insufficiently. The same authors even showed an improved survival in older women with high risk breast cancer following postmastectomy radiation therapy [30]. Similarly Hughes showed that irradiation of the breast in women ≥70 years treated with lumpectomy and tamoxifen for breast cancer resulted in an absolute reduction of locoregional recurrence of 5.9% [31]. She found the recurrence rate of 7% in the non-irradiated group acceptable and in view of the lack of benefit regarding breast cancer specific survival suggested omitting radiotherapy in this specific group of women.

Comorbidity in our population was relatively low, but women over the age of 80 were found to have a significantly lower recurrence rate when radiotherapy was added to their treatment regime. This suggests that even in the very elderly, standard locoregional treatment has clear advantages over surgical therapy only.

This study presents with some limitations inherent to a retrospective chart review. The number of very elderly patients with breast cancer in a single institution is small. A multi-institutional review, therefore, may provide more powerful recommendations. Also the rationale for omitting modalities of care was often not detailed in the reviewed charts. There was no controlling for patient choice or physician bias in treatment of the very elderly or for treatments offered for each stage of disease. A prospective, randomized trial controlling for comorbidity, stage at diagnosis and treatment administered will improve the general understanding of breast cancer in the very elderly. However, given the current state of practice in the United States, it is unlikely that prospective randomized studies will ever come to fruition.

Based on the currently available data however, we recommend that guidelines for breast cancer treatment in the very elderly mirror the currently accepted guidelines for younger women. In patients with significant comorbidity, risk-benefits analysis is warranted prior to initiating complete therapy.

Conflict of interest 

None declared.

Reviewer 

Professor Riccardo Audisio, Consultant Surgeon; Surgical Oncologist, Whiston Hospital, Department of Surgery, Warrington Road, Prescot, Merseyside L35 5DR, United Kingdom.