Factors Associated with Health-related Quality-of-life in Breast Cancer Survivors: Influence of the Type of Surgery

Abstract

Objective

To determine if health-related quality-of-life (QOL) differences existed between breast cancer (BC) survivors receiving mastectomy and those receiving breast-conserving treatment (BCT). Factors associated with QOL in long-term BC survivors were also identified.

Methods

One hundred patients who had previously undergone BC surgery and were alive without recurrence for >5 years were asked to answer the patient-administered questionnaires to assess their QOL (Functional Assessment of Cancer Therapy scale-Breast: FACT-B) and psychological distress (Hospital Anxiety and Depression Scale: HADS). Of them, 93 responded to the questionnaires affirmatively.

Results

Although none of the QOL scores were related to the surgical procedures, statistically significant relationships were found between age and the scores of FACT-General and social/family well-being (SWB), and between the educational status and scores of SWB in univariate analyses. There was no statistically significant relationship between psychological distress and each factor examined. In multivariate analyses, significant correlations were established between scores of the FACT-BC subscale (FACT-BCS) and the type of surgery and between those on the FACT SWB subscale and age at study or educational status. Namely, patients who had undergone BCT, younger patients and patients with higher educational background scored higher QOL.

Conclusions

Among the BC survivors, those who underwent BCT experienced significantly but slightly better QOL than those who received mastectomy in FACT-BCS assessments. Younger patients and patients with higher educational backgrounds experienced significantly better SWB.

breast cancer survivor, quality of life, type of surgery, breast-conserving treatment

INTRODUCTION

Since standard radical mastectomy was innovated by Halsted for breast cancer (BC) surgery, numerous randomized controlled trials (RCTs) to improve the prognosis of operable BC or health-related quality of life (QOL) of patients have been attempted (1–6). Comparative studies between standard radical mastectomy and modified radical mastectomy with RCTs have revealed no significant difference in prognosis (3,4). As less extensive surgery is expected to provide patients with better QOL, modified radical mastectomy has been preferred (to standard radical mastectomy) as the standard surgical procedure for early BC.

The subsequently introduced breast-conserving treatment (BCT), which was composes of wide local excision, axillary dissection and radiotherapy, was compared with modified radical mastectomy on RCTs with no apparent differences in prognosis (5,6). As such, BCT was accepted as one of the standard local treatments for early BC. Although BCT was expected to yield better QOL than mastectomy for BC patients, comparative studies verifying QOL results using the two methods proved otherwise (except that body image was superior to mastectomy even in RCTs) (7). The majority of cases examined indicated relatively short post-surgical periods prior to the study entry, e.g. <5 years (8,9). A few studies examined QOL in relation to the types of surgery in BC long-term survivors, and the results of them are relatively inconsistent.

Since BC is a common disease and prognosis of BC is relatively good in general, the number of BC survivors is very large and still now increasing. Therefore, QOL in BC survivors has been an important issue. The factors associated with QOL in BC survivors that have been reported are chemotherapy, medical condition, social support, income and so on (10). However, the results reported were also relatively inconsistent.

In the present study, we examined the QOL in BC patients who had survived 5 years or more without recurrence. Special attention was focused on the influence of the type of surgery on QOL.

PATIENTS AND METHODS

In this cross-sectional study, patients were enrolled based on the following criteria: (i) BC lumps of ≤3 cm in diameter by palpation at the time of surgery, (ii) underwent modified radical mastectomy or BCT at the National Hospital Organization Shikoku Cancer Center and (iii) survived without recurrence for ≥5 years. Post-operative consecutive BC patients (n = 100) who were visiting the outpatient clinic of one of the authors (S.O.) regularly were asked to complete the patient-administered questionnaires between July 2004 and March 2005 to assess QOL [Functional Assessment of Cancer Therapy scale-Breast: FACT-B (11)] and psychological distress [Hospital Anxiety and Depression Scale: HADS (12)]. The patients who received primary or secondary reconstruction were excluded from this study. All participants who received adjuvant endocrine treatment already finished it. The validated Japanese version of FACT-B is a 36-item questionnaire that measures both the 27-item general QOL associated with cancer (FACT-G) and the additional 9-item BC-related QOL [i.e. BC subscale (BCS) which asks questions regarding symptoms due to the disease or treatment such as shortness of breath and hair loss, and psychological aspects related to the disease and treatment such as worry over the effect of stress on the illness and feeling like a woman]. FACT-G has four subscales associated with physical (seven items), social/family (seven items), emotional (six items) and functional (seven items) well-being (13,14). The FACT system uses a five-point scale (i.e. 0–4), where a higher score indicates better QOL. HADS is a questionnaire used to assess the degree of depression and anxiety, and a four-point scale (i.e. 0–3) was employed, where the higher the score, the worse is the condition. An additional questionnaire inquiring the age, marital status, educational status, occupational status, household annual income and co-morbidity status of the patients was also used in this study. The questionnaires were administered and analyzed in a blind fashion; questionnaires were distributed to patients by the study supervisor (S.O.) in person, and the patients sent the completed questionnaires to the data center by mail subsequently. We planned to distribute the questionnaires to a total of 100 patients, of which half were patients treated with mastectomy and the other half were those with BCT. The questionnaires were handed to the patients who consecutively visited the outpatient office of S.O., fulfilled the criteria and consented to participate in the study. The number of patients treated with BCT was larger. Therefore, after the 50th BCT patient participated in the study, the questionnaires were handed to only patients treated with mastectomy. Patients were briefed on the blind treatment of questionnaires before obtaining their written consent to participate in the study.

Clinicopathological findings of the participating patients were then sent to the data center by the investigators. Office personnel at the data center subsequently pooled the data from both the participants and investigators in a blind fashion. The anonymously treated data were finally sent to the investigators for analysis.

The following exploratory variables correlated with QOL in BC patients: type of surgery (Mastectomy/BCT), age [60 years (close to the median) or younger/61 years or older], marital status (married/others), educational status (≤9/≥10 years), employment status (employed/unemployed), annual household income (≤5/5–10/>10 million yen), post-surgical period [≤84 (median)/≥85 months], pathological nodal status (negative/positive), adjuvant chemotherapy (no/yes), adjuvant endocrine therapy (no/yes) and co-morbidity such as hypertension and diabetes mellitus which is being treated (no/yes).

The relationships between the outcome variables and the explanatory variables were examined using both univariate and multivariate analyses. Univariate analyses included the one-way analysis of variance, χ² test and t-test. Linear regression analysis was used for multivariate analyses to examine the relationship of each of the outcome variables with all the explanatory variables. Since there were a few missing data in the completed questionnaires, no specific analytical techniques were applied to deal with the missing data.

The study protocol was approved by the National Hospital Organization Shikoku Cancer Center Institutional Review Board in May 2004.

RESULTS

Except for one case, who did not receive axillary dissection, all appropriately responded participants (93 patients) fulfilled the eligibility criteria. The median and mean age at the study and time-period after the definitive surgery were 58, 62.0 (range: 44–83) years old and 84 (range: 60–132) months, respectively. Apart from 47 patients (50.5%) who received mastectomies, the remaining patients underwent BCT. The characteristics of the participants are shown in Table 1. No patients were receiving either chemotherapy or endocrine therapy at the study.

Table 1.

Patient and tumor characteristics and treatments

Age at study	Median 58 years (range: 44–83 years)
Marital status
Married	77 (82.8%)
Others	15 (16.1%)
Unknown	1 (1.1%)
Educational status (years)
≤9	15 (16.1%)
≥10	74 (79.6%)
Unknown	4 (4.3%)
Occupational status
Having a job	39 (41.9%)
Having no job	52 (55.9%)
Unknown	2 (2.1%)
Household annual income (million yen)
≤5	52 (55.9%)
5.1–10	29 (31.2%)
>10	5 (5.4%)
Unknown	7 (7.5%)
Type of surgery
Mastectomy	47 (50.5%)
Breast-conserving treatment	46 (49.5%)
Duration after surgery	Median 84 months (range: 60–132)
Pathological nodal status
Negative	64 (68.8%)
Positive	28 (30.1%)
Unknown	1 (1.1%)
Adjuvant chemotherapy
No	47 (50.5%)
Yes	46 (49.5%)
Adjuvant endocrine therapy
No	28 (30.1%)
Yes	65 (69.9%)
Co-morbidity
No	60 (64.5%)
Yes	32 (34.4%)
Unknown	1 (1.1%)

On the basis of a comparison of the characteristics between the mastectomy and the BCT groups (Table 2), their backgrounds were not significantly different, albeit the number of patients with positive axillary lymph nodes tended to be more in the mastectomy group (P = 0.06). Among the patients treated with mastectomy, none received reconstruction surgery.

Table 2.

Comparison of the patient and tumor characteristics between the two treatment groups

	BCT (N = 46)	Mastectomy (N = 47)	P value
Age at study (mean)	58.9	61.0	0.29
Marital status
Married	38	39	0.56
Others	8	7
Unknown	0	1
Educational status (years)
≤9	7	8	0.59
≥10	36	38
Unknown	3	1
Occupational status
Having a job	20	19	0.96
Having no job	25	27
Unknown	1	1
Household annual income (million yen)
≤5	26	26	0.95
5.1–10	14	15
>10	2	3
Unknown	4	3
Time after surgery (mean)	85.4 months	85.7 months	0.95
Tumor size by palpation (mean)	2.0 cm	2.1 cm	0.69
Pathological nodal status
Negative	37	27	0.06
1–3	9	16
4–9	0	3
Unknown	0	1
Adjuvant chemotherapy
No	22	24	0.75
Yes	24	23
Adjuvant endocrine therapy
No	35	30	0.20
Yes	11	17
Co-morbidity
No	17	15	0.54
Yes	29	31
Unknown	0	1

BCT, breast-conserving treatment.

The results of univariate analyses demonstrated that none of the QOL scores were correlated with the surgical procedures (Table 3). Statistically significant relationships were found between the age and the FACT-G (P = 0.037) scores (mean scores: 87.6 for the patients of 60 years or younger vs. 81.5 for those of 61 years or older) and social/family well-being (SWB) (P = 0.001) (mean scores: 18.3 for the patients of 60 years or younger vs. 12.7 for those of 61 years or older) in univariate analyses; namely, the higher the age, the lower the QOL. Of particular interest was the significant relationship between educational status and SWB scores (P = 0.0001); namely, the higher the educational status, the higher the SWB scores (mean scores: 7.62 for the patients with education of ≤9 years vs. 17.3 for those with education of ≥10 years). No other statistically significant associations were found in univariate analyses between the exploratory variables and the QOL scores.

Table 3.

Scores of FACT and HADS by type of surgery

	BCT	Mastectomy	P value
FACT-G
Physical well-being	26.5 (2.05)	26.1 (2.87)	0.44
Social well-being	21.1 (8.71)	24.0 (10.5)	0.18
Emotional well-being	20.0 (2.95)	20.0 (3.42)	0.89
Functional well-being	22.1 (5.24)	23.3 (5.06)	0.32
FACT-G total	89.7 (13.1)	94.0 (14.0)	0.18
Breast cancer subscale	25.6 (4.38)	24.0 (4.37)	0.10
FACT-B total	115.4 (14.5)	118.2 (15.6)	0.42
HADS
Depression	4.03 (3.19)	3.18 (2.87)	0.18
Anxiety	5.03 (3.28)	5.30 (3.09)	0.68
Total	9.06 (8.47)	8.47 (5.44)	0.62

Values are mean (standard deviation). FACT-G, Functional Assessment of Cancer Therapy scale-General; FACT-B, Functional Assessment of Cancer Therapy scale-Breast; HADS, Hospital Anxiety and Depression Scale.

Multivariate analyses indicated significant correlations: (i) between FACT-BCS scores and type of surgery (P = 0.042) (Table 4) and (ii) between FACT SWB subscale scores and age at the study (P = 0.050) or educational status (P = 0.016) (Table 5). In other words, the BCT group had better FACT-BCS scores, whereas the younger patients and patients with higher educational backgrounds indicated higher FACT SWB scores. No other statistically significant relationships were identified in multivariate analyses.

Table 4.

Results of multivariate analysis in FACT-BCS

Variable	Reg. Coef.	SE	P value
Type of surgery (mastectomy/BCT)	−1.9	0.9	0.042
Age (61 years or older/60 years or younger)	1.5	1.1	0.178
Time after surgery (≥85/≤84 months)	−1.3	0.9	0.161
Lymph node status (positive/negative)	1.3	1.0	0.208
Chemotherapy (yes/no)	0.3	1.0	0.802
Endocrine therapy (yes/no)	−0.9	1.0	0.360
Marital status (married/others)	0.3	1.4	0.801
Educational status (≥10/≤9 years)	1.4	1.5	0.373
Employment status (employed/unemployed)	−1.6	1.0	0.091
Household income (>10/5–10/≤5 million yen)	−0.2	0.8	0.753
Co-morbidity (yes/no)	0.8	1.0	0.432

Reg. Coef., regression coefficient; SE, standard error.

Table 5.

Results of multivariate analysis in FACT-social/family subscale

Variable	Reg. Coef.	SE	P value
Type of surgery (mastectomy/BCT)	2.6	1.8	0.142
Age at study (61 years or older/60 years or younger)	−4.3	2.1	0.050
Time after surgery (≥85/≤84 months)	−0.6	1.7	0.726
Lymph node status (positive/negative)	1.5	2.0	0.437
Chemotherapy (yes/no)	0.6	1.9	0.771
Endocrine therapy (yes/no)	−0.1	1.9	0.976
Marital status (married/others)	0.2	2.5	0.921
Educational status (≥10/≤9 years)	7.3	3.0	0.016
Employment status (employed/unemployed)	−1.9	1.8	0.286
Household income (>10/5–10/≤5 million yen)	0.3	1.5	0.844
Co-morbidity (yes/no)	0.7	2.0	0.707

In order to examine what items in FACT-BCS show higher scores in the BCT group (vs. the mastectomy group) more specifically, we compared the mean scores of the respective FACT-BCS items with the type of surgery (Table 6). The higher original scores of B1, 2, 3, 5, 6, 7 and 8 indicate worse conditions; therefore, those scores should be reversed by subtracting the actual scores of those items from 4 to calculate FACT-BCS scores. In Table 6, mean scores of those reversed scores were shown for the scores of B1, 2, 3, 5, 6, 7 and 8. Namely, higher scores of all items in Table 6 represent better conditions. Only Item 2 (‘I am self-conscious about the way I dress’) showed a significantly (P = 0.006) higher (or better) mean score in the BCT than the mastectomy group.

Table 6.

Comparison of item scores of FACT-BCS between the two treatment groups

		BCT	Mastectomy	P value
B1	I have been short of breath	3.63 (0.70)	3.64 (0.58)	0.96
B2	I am self-conscious about the way I dress	3.17 (1.09)	2.43 (1.31)	0.006
B3	One or both of my arms are swollen or tender	3.31 (1.02)	3.24 (1.24)	0.79
B4	I feel sexually attractive	1.54 (1.22)	1.34 (1.24)	0.48
B5	I am bothered by hair loss	3.73 (0.67)	3.78 (0.61)	0.73
B6	I worry that other members of my family might someday get the same illness I have	2.70 (1.22)	2.19 (1.36)	0.08
B7	I worry about the effect of stress on my illness	2.37 (1.32)	2.33 (1.23)	0.89
B8	I am bothered by a change in weight	2.97 (1.19)	3.17 (1.10)	0.45
B9	I am able to feel like a woman	1.87 (1.12)	1.78 (1.28)	0.74

Values are mean (standard deviation).

DISCUSSION

The original objective of BCT was to conserve and avoid removal of the whole breast from women with operable BC in definitive surgery. Follow-up RCTs comparing BCT with mastectomy have not been able to show any difference in survival (5,6). Since QOL of the patients with BCT has generally been expected to be better than that of the patients with mastectomy, the interest of surgeons has moved to the patients' benefit on QOL. Numerous studies have been performed to determine whether BCT provides BC patients with better QOL. Although most of the studies were not RCTs where the surgical therapy preferred by the patient had been performed, some results were obtained from RCT-based studies. Of review articles that have extensively summarized those studies (7,15), the article by Kiebert et al. (7) was one of the most comprehensive articles. Since they included not only the results of RCTs but also those of non-randomized studies, we (16) further reviewed only RCTs from their article and complemented the findings of new RCTs that were reported after the article of Kiebert et al. (7). According to the results, differences in the respective QOL domains were not found between BCT and mastectomy, except for body image. However, in the majority of the studies quoted in the articles mentioned above, the investigators evaluated QOL of the patients who were followed for only several years after the definitive surgery. Note that patients who had undergone surgery several months before the study might have been receiving post-operative adjuvant systemic therapy and/or radiation therapy. Moreover, they might not have recovered from the psychological damage of having cancer and the fear of death. These factors could have affected QOL of the post-operative women. Therefore, it might not be appropriate to evaluate the post-operative QOL in patients on a short-term basis to differentiate the impact on QOL between BCT and mastectomy. However, only a few studies (17–19) examined QOL in relation to types of surgery in BC long-term survivors.

In the present study, there were no differences in the scores of total FACT-G, total FACT-B, BCT subscale and each domain of FACT-G, and HADS scores between the two surgical groups in univariate analyses. However, multivariate analyses showed a statistically significant difference between the surgical groups in BCT subscale scores (P = 0.042). Among the items of FACT BCT subscale, only the item regarding the way of dressing showed a statistically significant difference. Therefore, we would like to stress that there was not at least a large difference in QOL of Japanese BC survivors between the two treatment groups in case they selected the types of surgery by themselves, but BCT might have given slightly better QOL in terms of the ways they dressed. Our results were relatively similar to those of studies conducted by Dorval et al. (17) and Janni et al. (18) in which global QOL were similar between the two surgery groups, however, quite different from those of Arndt et al. (19) which showed statistically significant difference at 5 years between the two groups. Although we cannot explain the discrepancy among the studies, the differences in culture of the different countries and differences of numbers of participants among the studies might have affected the results.

Among factors other than the type of surgery, patients of younger ages at the study and those with higher educational backgrounds demonstrated significantly better FACT SWB scores in both univariate and multivariate analyses. Mols et al. (10) reviewed papers reporting on QOL among long-term BC survivors. They concluded that the most consistent factors associated with QOL in such patients were chemotherapy, medical condition, social support and income. Those factors were not identified as the predictive ones of QOL in our study. On the other hand, age at study and educational status were inconclusive factors in their review. Although we have any good idea to explain the discrepancy, the number of patients in our trial might have been too small to detect the small difference in QOL in those factors which were consistently associated with QOL in their review article, and the cultural difference between Japan and western countries might have affected the results here again.

Although we tried as far as possible to abolish factors biasing our results, certain issues remained unresolved in the present study: (i) the investigation was conducted as a cross-sectional and not a prospective longitudinal study, (ii) the type of surgery was performed according to patients' preference and not based on a randomized fashion, (iii) the sample size was relatively small and (iv) questionnaires were delivered to the patients directly by the study supervisor. With reference to factors (i) and (ii), characteristics of the two treatment groups were similar and post-operative evaluation periods of QOL approximated well in both the groups. But selection bias was inevitable in non-randomized studies such as this study. Therefore, the present results should be regarded as those of a QOL study in patients who received surgical treatments they preferred and did not experience recurrences. As for factor (iii), further studies are warranted to confirm the present findings. Factor (iv) might have forced the patients to provide inflated scores (possibly better than their own actual conditions). However, the patients were assured that the completed questionnaires would be analyzed in an anonymous and independent fashion. Moreover, if this factor had affected the results, the scores of both surgical groups would have been inflated in the same fashion.

In conclusion, among the BC survivors, although there was not at least a large difference in QOL between the two treatment groups, those who underwent BCT experienced significantly but slightly better QOL than those who received mastectomy in FACT-BCS assessments, especially in terms of the ways they dressed. Therefore, we would like to recommend clinicians not to say that BCT is much better than mastectomy in terms of QOL, but explain the merits and demerits of both surgical procedures to patients and ask them to decide which is better for them. And younger patients and patients with higher educational backgrounds experienced significantly better SWB.

Funding

This work was supported by a Grant-in-Aid for research on cancer treatment from the Ministry of Health, Labor, and Welfare of Japan (13-10).

Conflict of interest statement

None declared.