In this controlled postdiagnosis study, the authors examined various aspects of body image of breast cancer survivors in cross-sectional and longitudinal designs.
In this controlled postdiagnosis study, the authors examined various aspects of body image of breast cancer survivors in cross-sectional and longitudinal designs.
In 2004 and 2007 the Body Image Scale (BIS) was completed by the same 248 disease-free women who had been treated for stage II and III breast cancer between 1998 and 2002. “Poorer” body image was defined as greater than the 70th percentile (N = 76 women) of the BIS scores in contrast to “better” body image (N = 172 women). Breast cancer survivors were examined clinically in 2004, and their BIS scores were compared with the scores from an age-matched group of women from the general population.
In this cross-sectional study, poorer body image in 2004 was associated significantly with modified radical mastectomy, undergoing or planning to undergo breast-reconstructive surgery, a change in clothing, poor physical and mental health, chronic fatigue, and reduced quality of life (QoL). In univariate analyses, most of these factors and manually planned radiotherapy were significant predictors of poorer body image in 2007. In multivariate analyses, manually planned radiotherapy, poor physical QoL and high BIS score in 2004 remained independent predictors of a poorer body image in 2007. Body image ratings were relatively stable from 2004 to 2007. Twenty-one percent of breast cancer survivors reported body image dissatisfaction, similar to the proportion of dissatisfaction in controls.
In this cross-sectional analysis, body image in breast cancer survivors was associated with the types of surgery and radiotherapy and with mental distress, reduced health, and impaired QoL. Body image ratings were relatively stable over time, and the antecedent body image score was a strong predictor of body image at follow-up. Body image in breast cancer survivors differed very little from that in controls. Cancer 2010. © 2010 American Cancer Society.
Body image is defined commonly as a subjective picture of an individual's own physical appearance established by self-observation and by noting the reactions of others.1 However, body image also includes an attitude of satisfaction or dissatisfaction with one's body that varies by 2 factors: investment in appearance and self-evaluation.2 Investment in appearance concerns an individual's view on the importance of his or her appearance. Self-evaluation relates to cultural ideals for physical appearance and beauty and the discrepancy between perceived body image and these ideals. Body image is linked closely to identity, self-esteem, attractiveness, sexual functioning, and social relationships.3 Psychological research has documented that body image problems are associated with poor self-esteem, social anxiety, self-consciousness, and depressive symptoms.4
Women who are treated for breast cancer are exposed to marked changes in their physical appearance, such as loss or disfigurement of 1 or both breasts, scars from surgery, and skin changes related to radiotherapy. Furthermore, systemic treatments with chemotherapy or hormones often lead to weight increase.5, 6 These changes are related intimately to physical appearance and body image, and an obvious question is how breast cancer and its treatment affect the body image of women. The main answers to date are that mastectomy (rather than breast-conserving surgery), young age at diagnosis, poor mental (but not physical) health-related quality of life (QoL), and psychological distress are associated significantly with poor body image.7-11 One study indicated that better cosmetic appearance of the breasts was associated significantly with less mental distress.12 In contrast, in an experimental longitudinal study, no significant differences were observed in psychological distress or body image between groups that did or did not undergo postmastectomy plastic surgery.13
In their review, Helms et al14 stated that the existing literature on body image and breast cancer patients has limitations, such as small to modest sample sizes, use of retrospective design, lack of control groups, and various selection biases. In the current study of breast cancer survivors (BCSs) 4 years and 7 years after diagnosis, we attempted to overcome some of these limitations by having a considerable sample size, a longitudinal design, and a control group sampled from the general population. We explored 3 research questions: 1) What are the characteristics of BCSs who report poorer body image compared with those who report better body image? 2) Which demographic, psychosocial, and treatment variables evaluated 4 years postdiagnosis predict poorer body image 3 years later? 3) Are there significant differences in the proportions of body image dissatisfaction between BCSs and age-matched women from the general population?
The current study group comprised a subgroup of BCSs who participated twice in a longitudinal survey that addressed long-term morbidity after treatment.15 BCSs who received postoperative locoregional radiotherapy for stage II and III breast cancer between 1998 and 2002 at the Norwegian Radium Hospital were surveyed first in 2004. The eligibility criteria for the survey and the sampling of patients have been described previously.15 In 2004, 318 tumor-free BCSs delivered valid questionnaires and had an outpatient examination. The 2007 survey invited 310 of the BCSs from 2004 (8 women had died or had developed recurrent disease since 2004). The 248 women who completed the mailed questionnaires in both 2004 and 2007 were eligible for the current study.
Treatment was based on the guidelines of the Norwegian Breast Cancer Group at that time (available at: http://www.nbcg.no/index.html; accessed July 6, 2009) and has been described in detail previously.15 The women underwent either modified radical mastectomy (MRM) (N = 184; 74%) or breast-conserving surgery (N = 64, 26%) and axillary lymph node dissection at level I/II. To our knowledge, the surgical approach was determined only by the surgeons. Before 2000, postoperative radiotherapy (50 grays) was based on a standardized, manually planned field arrangement that consisted of an anterior supraclavicular/axillary field, a posterior field that covered the axilla, 2 tangential fields, and an electron field that covered the parasternal lymph nodes. From 2000 onward, computed tomography (CT) planning with the planning system Helax-TMS (Helax AB, Uppsala, Sweden) was used based on transverse CT images. Manually planned radiotherapy was received by 87 BCSs (35%), and CT-planned radiotherapy was received by 161 BCSs (65%). Adjuvant chemotherapy and/or tamoxifen were administered according to national guidelines at that time. Tumor stage and the number of axillary lymph nodes removed were noted.
Information on sociodemographics, lifestyle, and physical and mental health variables was obtained by using a questionnaire. Relationship status was dichotomized as paired (married or cohabiting) versus nonpaired (separated, divorced, or widowed). The number of children was grouped into ≤1 child or >1 child. Level of education was categorized as ≤10 years, 11 to 12 years, or ≥13 years of basic education. Work status was grouped according to women in paid work (full time, part time, or self-employed) versus those not in paid work (unemployed, pensioned, on social benefits, students, or homemakers).
Body mass index (BMI) was dichotomized as <27 kg/m2 versus ≥27 kg/m2. Data on weight before chemotherapy and at the time of the surveys enabled the assessment of weight changes over time. Self-rated health was scored by 1 item on a 4-point Likert scale that ranged from “very good” to “poor,” and the results were dichotomized into categories of “very good/good” versus “moderate/poor.” Regular use of medication was noted if drugs were used regularly during the last 12 months and concerned analgesics and psychotropics (hypnotics, tranquilizers, or antidepressants). The level of physical activity was categorized into “minimal activity” versus “moderate or more activity” as defined by published algorithms.16
At the outpatient examination, an oncologist and experienced cancer nurses examined the patients for dyspigmentation, telangiectasis, and fibrosis within the treated breast area. “Some,” “moderate,” or “much” dyspigmentation and fibrosis were rated as present, and telangiectasis was rated as present or absent. Any performed breast reconstruction was recorded, and plans for future plastic surgery were noted. Whether patients had changed their clothing after treatment also was recorded. The presence of lymphedema was defined according to a published definition used by our group.17
The Body Image Scale (BIS) is a 10-item, self-rating scale that was developed to ascertain changes in the body image of cancer patients.18 Five BIS items concern general body image issues: feeling self-conscious, dissatisfied when dressed, difficulty looking at yourself naked, avoid others because of appearance, and dissatisfied with body. The other 5 BIS items concern body image in relation to the cancer experience: less physically attractive, less feminine, less sexually attractive, body less whole, and dissatisfied with scar.
The time frame for the BIS is the past week. Each item is scored on a 4-point Likert scale as follows: 0, not at all; 1, a little; 2, quite a bit; and 3, very much. Higher scores represent poorer body image. Because no cutoff value for body image problems has been defined for the BIS, we defined those greater than the 70th percentile of the BIS score in 2004 as the poorer body image group (BIS score ≥8) compared with the other BCSs who were defined as the better body image group (BIS score <8). These definitions also were used for the 2007 BIS scores.
Each BIS item score also was dichotomized into “dissatisfied” (score 2-3) or satisfied (score 0-1).19 Internal consistency of the BIS had a Cronbach α = .93 in our sample.
The Fatigue Questionnaire20 consists of 13 items, and 11 of those items assess the prevalence and intensity of fatigue symptoms compared with the last time the respondent felt well. Seven items concern physical fatigue, and 4 items relate to mental fatigue. Higher scores suggest more fatigue. Cases of chronic fatigue were defined according to Fukuda et al21 as a sum score ≥4 on dichotomized Fatigue Questionnaire item scores and a symptom duration ≥6 months. The internal consistency was between α = .85 for mental fatigue and α = .92 for physical fatigue.
The Medical Outcomes Study 36-item Short Form Health Survey (SF-36) is a generic QoL instrument that consists of 36 items and assesses 8 dimensions of physical and mental QoL.22 The dimensional scores may be summarized into a Physical Component Summary (PCS) score and a Mental Component Summary (MCS) score. For the current study, these scores were T-transformed so that the Norwegian general population mean score was set at 50.23
The European Organization for Research and Treatment of Cancer QLQ-BR23 instrument (BR23) is a 23-item, breast cancer-specific module of 5 QoL dimensions; of those, we included the dimensions sexuality, arm problems, and breast symptoms in the current study. The items are rated on a 4-point Likert scale from 1 (not at all) to 4 (very much). The BR23 has been tested in an American sample with adequate internal consistencies.24 The current internal consistencies were α = .86 for sexuality, α = .75 for arm problems, and α = .80 for breast symptoms.
The Hospital Anxiety and Depression Scale (HADS) consists of 14 items, with 7 items that assess symptoms of anxiety (HADS-A) and 7 items that assess depression (HADS-D). On the basis of the literature, HADS-defined anxiety disorder or depression were defined by a score ≥8 on the HADS-A and HADS-D, respectively.25 The internal consistency was α = .88 for HADS-A and α = .83 for HADS-D.
In 2004, the NHR addressed 3500 women ages 20 to 79 years in the general population with a questionnaire that, among other scales, included the 5 BIS items concerning general body image applicable to women without cancer. The response rate was 41%; and, among the 1435 responders, 2 age-matched controls for each BCS in our sample were used as normative controls (N = 496). Internal consistency was α = .84.
To have an adequate sample size for BCSs with poorer body image, the group was defined by a BIS score greater than the 70th percentile (see above). Continuous variables were analyzed with t tests, and categorical variables were analyzed with chi-square tests. In case of skewed distributions, nonparametric tests were applied. Statistically significant differences were tested for clinical significance according to their effect size (ES) with a cutoff of ES ≥0.40.26, 27
Internal consistencies of scales were determined by Cronbach coefficient (α), correlations with Spearman coefficient (σ), and concordance between groups with the Cohen coefficient (κ). Univariate and multivariate logistic regression analyses were used to explore the associations between demographic, lifestyle, and cancer-related variables measured in 2004 as independent variables and poorer versus better body image in 2007 as a dependent variable. Because of the limited number of women with poorer body image in 2007 (n = 67) only the most relevant independent variables from the univariate analysis were included in the multivariate analysis. The strength of the associations was expressed as odds ratios (ORs) with 95% confidence intervals (CIs). For these analyses, we used the software package SPSS for Windows, version 15.0 (SPSS, Inc., Chicago, Ill). The level of significance was set at P < .05, and all tests were 2-sided.
This study was approved by the Regional Ethical Committee of Health Region South of Norway and by the National Data Inspectorate. All patients provided written informed consent.
When the 70 nonresponders in 2004 were compared with the 248 responders in 2004, no significant differences were observed except that the HADS-A score was higher in nonresponders (P = .04; ES = 0.28; data not shown).
Seventy-six BCSs (31%) had poorer body image in 2004. The groups with poorer and better body image did not differ significantly on demographic variables (Table 1). Concerning cancer-related variables, a greater proportion of women who had a poorer body image had undergone MRM and had undergone or wanted to undergo breast-reconstructive surgery. Furthermore, women who had a poorer body image more frequently had developed fibrosis in the treatment area and had changed their clothing to a considerable extent. All of these differences also were significant clinically, except for development of fibrosis. The finding that a greater proportion of women who had received manually planned radiotherapy had a poorer body image indicated borderline significance. No significant group differences were observed concerning tumor stage, the number of lymph nodes removed, the presence of lymphedema, treatment with chemotherapy, hormone use, age at diagnosis, or age at survey. Women who had undergone both MRM and manually defined radiotherapy were significantly more common in the poorer body image group.
|Variable||No. of Survivors (%)||P||ES|
|Poorer Body Image Score (n=76)||Better Body Image Score (n=172)|
|Age: Mean±SD, y|
|Follow-up: Mean±SD, y||4.2±0.9||4.1±0.9||.27|
|Partner||62 (82)||136 (80)|
|No partner||14 (18)||35 (20)|
|No. of children||.24|
|0-1||12 (16)||37 (22)|
|2-7||64 (84)||129 (78)|
|Level of education, y||.18|
|≤10||17 (22)||33 (19)|
|11-12||37 (49)||67 (39)|
|≥13||22 (29)||70 (41)|
|Paid work||39 (51)||97 (57)|
|No paid work||37 (49)||73 (43)|
|Type of surgery||<.001||0.58|
|Mastectomy||68 (90)||116 (67)|
|Breast-conserving surgery||8 (10)||56 (33)|
|II||65 (87)||150 (88)|
|III||10 (13)||20 (12)|
|Menopausal status at surgery||.93|
|Premenopausal||38 (50)||84 (49)|
|Postmenopausal||38 (50)||86 (51)|
|Received chemotherapy||61 (80)||137 (80)||.91|
|Received hormones||70 (92)||147 (86)||.15|
|CT-based||43 (57)||118 (69)|
|Manually defined||33 (43)||54 (31)|
|No. of lymph nodes removed||.26|
|N0||4 (5)||7 (4)|
|N1: 1-3||50 (66)||101 (59)|
|N2: 4-9||17 (22)||58 (34)|
|N3: ≥10||5 (7)||6 (3)|
|Lymphedema present||16 (21)||23 (13)||.13|
|Underwent breast reconstruction||26 (34)||23 (14)||<.001||0.48|
|Wants breast reconstruction||18 (25)||12 (7)||<.001||0.51|
|Changed clothing||33 (43)||26 (15)||<.001||0.47|
|Skin changes in breast area||.96|
|Dyspigmentation||17 (24)||39 (24)|
|Teleangiectasies||50 (70)||100 (60)||.12|
|Fibrosis||33 (47)||50 (29)||.02||0.37|
|Mastectomy plus manually defined radiotherapy||32 (42)||42 (24)||.005||0.39|
A significantly greater proportion of BCSs with poorer body image reported poorer health, use of analgesics and psychotropics, diagnoses of probable anxiety disorder or depression, and chronic fatigue compared with survivors in the better body image group (Table 2). These differences reached clinical significance, except for the use of medication. Women in the poorer body image group had significantly lower mean scores on both the generic and disease-specific QoL scales, but only the generic scores reached clinical significance (Table 2).
|Health-Related Variable||No. of Survivors (%)||P||ES|
|Poorer Body Image Score, n=76||Better Body Image Score, n=172|
|Very good/good||37 (49)||135 (79)|
|Moderate/poor||39 (51)||36 (21)|
|BMI ≥27 kg/m2||33 (48)||53 (36)||.10|
|Minimal||5 (7)||17 (10)|
|Moderate or greater||71 (93)||155 (90)|
|Regular use of medication|
|Analgesics||19 (25)||22 (13)||.02||0.31|
|Psychotropics||26 (34)||31 (18)||.005||0.37|
|Depression||19 (25)||10 (6)||<.001||0.55|
|Anxiety||40 (53)||39 (29)||<.001||0.49|
|Chronic fatigue||34 (45)||44 (26)||.003||0.40|
|Generic health-related QoL score, mean±SD|
|Disease-related health-related QoL, mean±SDa|
Sixty-seven BCSs (27%) had a poorer body image in 2007. Among those 76 survivors who had a poorer body image in 2004, 52 survivors (68%) also had a poorer body image in 2007, whereas 24 survivors (32%) had changed to a better body image. Among the 172 BCSs who had a better body image in 2004, 15 survivors (9%) had changed to a poorer body image in 2007, whereas 157 survivors (91%) retained a better body image over time. The κ value for stability of body image was 0.62. The correlations between the 2004 and 2007 BIS summary scores had a coefficient of σ = 0.78.
The mean body image scores for survivors who had received manually planned radiotherapy were significantly poorer than the scores for survivors who had received CT planned radiotherapy both in 2004 and in 2007 (data not shown). In univariate analyses, most of the independent variables that differed significantly between the poorer and better body image groups in 2004 were significant predictors of a poorer body image in 2007 in addition to manually planned radiotherapy (Table 3). A higher BIS score in 2004 also significantly predicted a poorer body image in 2007.
|Variable||Univariate Analysis||Multivariate Analysis|
|OR||95% CI||P||OR||95% CI||P|
|Mastectomy (referent: lumpectomy)||2.81||1.30-6.08||.008||0.53||0.13-2.08||.36|
|Manually based radiotherapy (referent: CT-based)||2.49||1.40-4.43||.002||3.66||1.43-9.38||.007|
|Want breast reconstruction||3.85||1.76-8.44||.001||0.87||0.48-1.60||.66|
|Moderate/poor self-rated health (referent: good/very good)||3.19||1.77-5.76||<.001||0.59||0.18-2.00||.40|
|Regular use of analgesics||3.68||1.84-7.36||<.001||—||—||—|
|Regular use of psychotropics||2.50||1.34-4.68||.004||—||—||—|
|Generic health-related QoL|
|Diseased-related, health-related QoL|
|BIS 2004 score||1.44||1.31-1.58||<.001||1.48||1.30-1.67||<.001|
In the multivariate logistic regression analysis, increasing BIS score, decreasing PCS score, and manually planned radiotherapy in 2004 remained significant predictors of a poorer body image in 2007 (Table 3). The weight changes from the postoperative state to 2004 were minimal (median, 73 kg vs 72 kg, respectively) as were the changes from 2004 to 2007 (median, 72 kg vs 71 kg, respectively).
Compared with controls, a significantly higher proportion of BCSs in 2004 were “dissatisfied” with regard to the general body image items (feeling self-conscious and being dissatisfied when dressed) compared with controls (Table 4). A significantly lower proportion of BCSs had no “dissatisfied” scores. However, none of these differences reached clinical significance. The proportions with “dissatisfied” scores are illustrated in Figure 1.
|BIS General Body Image Items||No. of Women (%)||P||ES|
|2004 BCS, n=248||Controls, n=496|
|Item 1. Feeling self-conscious||24 (9)||19 (4)||.002||0.24|
|Item 3. Dissatisfied when dressed||20 (8)||16 (3)||.004||0.22|
|Item 5. Difficult looking at self naked||24 (10)||32 (7)||.11|
|Item 7. Avoid people because of appearance||8 (3)||7 (1)||.10|
|Item 9. Dissatisfied with body||38 (15)||56 (11)||.12|
|No. of “dissatisfied” ratings||.02||0.19|
|None||197 (79)||426 (86)|
|≥1||51 (21)||70 (14)|
The cross-sectional examination in 2004 revealed that poorer body image in BCSs was associated significantly with MRM, breast-reconstructive surgery, a wish for breast reconstruction, having a change in clothing, and the development of fibrosis in the treatment area. Poorer body image also was associated with poorer self-rated health, chronic fatigue and mental distress, and poorer generic and disease-related QoL in univariate analyses. Most of these variables also significantly predicted poorer body image 3 years later. In multivariate analysis, however, increasing BIS score, reduced PCS score, and treatment with manually planned radiotherapy in 2004 remained significant predictors of a poorer body image at follow-up. The persistence of a poorer body image over 3 years was considerable. The differences in body image satisfaction between BCSs and controls were statistically (but not clinically) significant. This finding has a positive aspect; ie, that, for practical purposes, the proportion of BCSs with body image dissatisfaction was not higher than that among controls. However, the controls did not have breast cancer, and the body image dissatisfaction among BCSs must be taken seriously.
Our 2004 survey of tumor-free BCSs at a mean of 4 years after primary treatment confirmed previous findings that poorer body image is associated with MRM, and mental distress. In contrast to those studies, we observed that physical QoL rather than mental QoL was associated significantly with a poorer body image. Associations with current age or age at diagnosis were not confirmed in our sample, possibly because of the low proportion of younger BCSs in our sample (only 25% were aged <50 years in 2004).
Surprisingly, a significantly greater proportion of women who underwent breast reconstruction were among the BCSs with poorer body image. This finding may suggest that such surgery hardly improved body image in our sample. That physical appearance is not the only decisive factor regarding body image also is supported by our findings that 42 of 74 BCSs (58%) who were treated with both MRM and manually planned radiotherapy had a better body image. Such radiotherapy was associated with significantly more skin changes than CT-based radiotherapy (data not shown). However, the type of radiotherapy had only a borderline significant difference between a poorer body and a better body image in 2004.
Our finding that a poorer body image was associated significantly with a wish for breast-reconstruction surgery and a change in clothing seems obvious, and both of these variables also were predictive of a poorer body image at follow-up in univariate analyses. However, taken together with the poorer body image of many BCSs who underwent breast reconstruction, these findings support current theories that an investment in appearance is of relevance for body image. In multivariate analyses, the type of radiotherapy was a significant predictor of a poorer body image at follow-up, whereas the type of surgery was not.
To our knowledge, this is the first prospective naturalistic study of body image in BCSs. The body image did not change much during our 3-year observation period, when women are close to becoming long-term survivors (≥5 years). This may be related to the minimal weight changes in our sample during the observation period. An alternative explanation is the trait-like character of body image with stability over time. Body image is a part of the personality that is not easily modified after it has become established, not even by treatment for breast cancer.
The low proportions of body image dissatisfaction and the lack of clinically significant differences compared with controls indicated that most BCSs have a good body image. Those who have a poorer body image usually also have problems in several other areas like fatigue, mental distress, and QoL, and they are in need of further evaluation and care.
Our study has the strengths of considerable sample sizes, both longitudinal and cross-sectional design, use of an established body image instrument, and hardly any differences between responders and nonresponders to the 2004 survey. Limitations concern lack of data on partner reactions to breast cancer and pretreatment data.
The construct of body image is complex, and no strong consensus exists about the content of the term.28 The BIS was developed from pragmatic considerations19 and was not driven by theories or models. Therefore, our study does not expand on the theory of body image. This might be regarded as a weakness; however, we believe that our results have practical implications and may contribute to a deeper understanding of body image problems in long-term BCSs. For clinical management, we have demonstrated that the modalities of surgery and radiotherapy have a significant influence on the body image of BCSs.
In conclusion, our cross-sectional analysis of BCSs indicated that MRM and local body changes after treatment as well as mental distress, self-rated health, and QoL were associated significantly with a poorer body image. However, body image among BCSs differed very little from that of a group of age-matched controls without cancer. The body image rating was relatively stable over 3 years, and the antecedent BIS score was a strong predictor of a poorer body image at follow-up along with the type of radiotherapy and physical QoL.
The authors made no disclosures.