Fax: (416) 978-8222
The impact of having a sister diagnosed with breast cancer on cancer-related distress and breast cancer risk perception†
Article first published online: 7 FEB 2013
Copyright © 2013 American Cancer Society
Volume 119, Issue 9, pages 1722–1728, 1 May 2013
How to Cite
Metcalfe, K. A., Quan, M.-L., Eisen, A., Cil, T., Sun, P. and Narod, S. A. (2013), The impact of having a sister diagnosed with breast cancer on cancer-related distress and breast cancer risk perception. Cancer, 119: 1722–1728. doi: 10.1002/cncr.27924
We wish to acknowledge the Canadian Breast Cancer Foundation (Ontario Chapter) for funding this study.
- Issue published online: 22 APR 2013
- Article first published online: 7 FEB 2013
- Manuscript Accepted: 15 OCT 2012
- Manuscript Revised: 9 OCT 2012
- Manuscript Received: 21 JUN 2012
- breast cancer;
- cancer risk perception
A family history of breast cancer has been shown to affect psychosocial functioning. However, the majority of research has focused on the daughters of patients with breast cancer and families with multiple relatives with the disease. The purpose of the current study was to examine cancer-related distress and breast cancer risk perception, and further examine the predictors of these outcomes, in the sisters of newly diagnosed patients with breast cancer without a previous family history of the disease.
Sisters of newly diagnosed index breast cancer patients were identified and asked to complete a study-specific questionnaire (demographics and cancer risk perception) and the Impact of Events Scale. Pathological information was abstracted from the medical chart for the index breast cancer patients.
A total of 205 sisters completed the questionnaires. The mean time between breast cancer diagnosis and the sisters' completion of the questionnaire was 9.8 months. Approximately one-half of the women scored in the moderate or severe distress range. The most significant predictor of cancer-related distress was perceived lifetime breast cancer risk (P = .04). Women with a lifetime risk of breast cancer > 20% were more than twice as likely to have moderate or severe distress compared with those with a lifetime risk of < 20%.
Cancer-related distress is high in the sisters of newly diagnosed patients with breast cancer in whom there is no other family history of breast cancer. Specifically, women with a perceived lifetime risk of breast cancer of > 20% experienced the highest levels of distress. Future interventions that target this group should be considered. Cancer 2013. © 2013 American Cancer Society.
According to the Canadian Cancer Society, approximately 22,000 women in Canada are diagnosed with breast cancer annually.1 Women diagnosed with breast cancer, as well as their relatives, experience high levels of distress after the diagnosis. Interestingly, distress levels in relatives are comparable to those observed in the patients themselves, suggesting that a breast cancer diagnosis has significant psychosocial implications for relatives.2 Distress in relatives has been associated with decreased natural cytotoxic activity and elevated levels of stress hormones.3
The majority of research examining the psychosocial implications of a breast cancer diagnosis in relatives has focused on daughters. In a qualitative study of the adult daughters of newly diagnosed patients with breast cancer, the narratives documented that the mother's cancer diagnosis subjected the daughters to a period of crisis fraught with severe emotional distress and life-and-death concerns.4 They also described a heightened sense of vulnerability. In a study by Mosher et al of adult daughters of women who were newly diagnosed with breast cancer, 19% of women fit the criteria for posttraumatic stress disorder and symptom severity was associated with the mother's stage of disease at the time of diagnosis.5
A family history of breast cancer has been shown to effect psychosocial functioning. Previous research has shown that both cancer-related distress and overall distress are elevated in women with a family history of breast cancer.6-11 However, most studies have focused on high-risk families in which multiple family members are affected with breast cancer. To the best of our knowledge, there has been very little research that has examined the psychosocial implications of a recent breast cancer diagnosis in families in which this is the first diagnosis of the disease. Eight of 9 women who develop breast cancer do not have an affected mother, sister, or daughter.12 For the majority of times a woman is diagnosed with breast cancer, it is the first time that sisters now realize that they are at an increased risk of developing the disease.
The purpose of the current study was to examine cancer-related distress and breast cancer risk perception, and to evaluate the predictors of these outcomes, in the sisters of patients newly diagnosed with breast cancer without a previous family history of the disease.
MATERIALS AND METHODS
All patients who were newly diagnosed with invasive breast cancer and who received surgical treatment at 2 academic health sciences centers in Toronto, Ontario, Canada were identified and approached for participation. All patients who were interested in participation were screened for eligibility criteria. This included being able to read and understand English, age > 25 years, a diagnosis of invasive breast cancer (not in situ) within the last 6 months, having at least 1 sister, and meeting family history criteria (no first-degree relative with breast cancer). If the patient met eligibility criteria, a detailed family history was obtained and eligible sisters were identified. These included sisters who could read and understand English, were aged ≥ 25 years, and had had no previous personal history of cancer. The patient was asked to contact each eligible sister and obtain consent for the study coordinator to contact them and explain the study. At the time of receipt of the contact information from the patient, each consenting sister was contacted and a study package was mailed to each subject for completion.
Medical Chart Review
Information concerning the patient was obtained through a systematic medical chart review (AJCC staging system). The review included recording data on patient age, the date of surgery, the type of surgery, and pathological information (including size, lymph node status, stage, and grade).
This questionnaire was designed specifically for the current study and was based on questionnaires that we have used in previous research. Items in the questionnaire included demographic information (age, marital status, education, etc) and information necessary to generate breast cancer risk estimates (including age at first period, number of biopsies, results of biopsies, age at first live birth, ethnicity, and previous diagnoses of ductal carcinoma in situ or lobular carcinoma in situ). We asked women about their perception of their lifetime risk of breast cancer after their sister's diagnosis using a Likert-type scale from 0% to 100%.
Impact of Event Scale
The Impact of Event Scale (IES)13 is a self-report measure that was designed to measure current subjective distress in relation to a specific stressor. For the current study, the stressor was identified as “having a sister with breast cancer.” It is based on a 2- factor intrusion and avoidance model of reactions to stressful life events. It measures the frequency of intrusive and avoidant phenomena. The scale consists of 15 items (7 intrusion items and 8 avoidance items). The IES allows for the calculation of a total score (with a possible range of 0-75), and separate intrusion and avoidance subscale scores. The Cronbach alpha is .78 for intrusion and .82 for avoidance. Scores for the total IES can be divided into the subclinical range (0-8), mild range (9-25), moderate range (26-43), and severe range (≥ 44).
Breast Cancer Risk Estimation
For each woman, the breast cancer risk was calculated with the Tyrer-Cuzick model14 using the International Breast Cancer Intervention Study (IBIS) program. This model uses the Bayes theorem to calculate the breast cancer probability and then refines the calculation by maximum likelihood estimates and incorporates risk factors such as age at menopause and menarche, weight, height, age, use of hormone replacement therapy, and previous benign breast biopsies. When compared with the Gail, Claus, and Ford models, the Tyrer-Cuzick model has been shown to be the most consistently accurate for the prediction of breast cancer15, 16 and BRCA (breast cancer gene) mutation carrier status.16 Specifically, with regard to breast cancer risk in women with 1 first-degree relative with breast cancer (consistent with those women in the current study), the Tyrer-Cuzick model has been shown to be accurate in predicting breast cancer risk whereas the Gail, Claus, and Ford models have all been shown to underestimate risk in this particular subgroup of women.17
Descriptive statistics (including frequencies and means) were generated for each of the study variables. Predictor variables were grouped and analysis of variance was used to compare the mean perceived lifetime risk of breast cancer between groups. The dependent variable total cancer-related distress was dichotomized into low total distress (IES total score of < 20) and high total distress (IES total score of ≤ 20) based on clinical cutoff scores. Predictors of high and low total distress were examined using logistic regression models. Statistical analyses were performed using SAS statistical software (version 9.1.3; SAS Institute, Inc, Cary, NC).
In total, 2084 women with breast cancer (index breast cancer patients) were identified, 1873 of whom were subsequently found to be ineligible (348 women had a family history of breast cancer; 145 women had a language barrier; 323 women had a breast cancer recurrence; 16 women were too young or had sisters who were too young to participate; 84 women had sisters who were residing out of the country; 200 women did not have a sister; 143 women had noninvasive breast cancer; 507 women were diagnosed with breast cancer before March 2007 or after August 2010; 21 women had other reasons for not participating, including patients who were adopted, those who had just received their diagnosis and were overwhelmed, and those who were not well enough to participate in the study; and 86 women refused participation). Of the 211 remaining patients with breast cancer, 152 agreed to participate and 236 sisters were identified and contacted about their participation. A total of 213 of the 236 sisters agreed to participate, and of these 213 sisters, 205 (96.2%) (from 147 families) completed the questionnaires. The mean age of the sisters was 50.1 years (range, 24 years-78 years). Characteristics of the sisters are presented in Table 1. One hundred fifty-six of the women identified themselves as white, 138 women had at least a college education, and 124 were married or in a common-law relationship.
|Patients With Breast Cancer|
|Mean age at diagnosis (range), y||50.1 (24-78)|
|Married/common law||124 (60.5)|
|Missing data||3 (1.5%)|
The index breast cancer patients were diagnosed with breast cancer between March 2007 and August 2010. Their characteristics are presented in Table 1. The mean time between breast cancer diagnosis and completion of the sisters' questionnaire was 9.8 months (range, 0 months-41 months). A total of 153 sisters (74.6%) completed the questionnaires within 12 months of the breast cancer diagnosis.
Breast Cancer Risk
Sisters were asked about their perceived lifetime risk of breast cancer after the diagnosis of the index breast cancer patient. The mean self-reported current breast cancer risk after diagnosis of the patient was 34.6% (range, 0%-100%). Age of the sister and model-generated (IBIS) breast cancer risk were found to significantly predict current perceived breast cancer risk (Table 2). Women aged < 50 years had a higher perceived lifetime risk of breast cancer (39.5%) compared with those aged ≥50 years (30.6%) (P = .02). Those women with a higher (≥ 10%) model-generated lifetime breast cancer risk estimated their current lifetime risk of breast cancer to be higher than those with a lower lifetime risk of breast cancer (< 10%) (P = .0008). Factors associated with the index breast cancer patient (including age at diagnosis, stage of breast cancer, and time since diagnosis) were not predictive of the perceived risk of breast cancer in the sisters.
|Factors||Variables (Frequencies) N=205||Mean Perceived BC Risk (Range) N=130||Pa|
|20-49 (92)||39.5 (0-100)||.02|
|50-80 (113)||30.6 (0-100)|
|Single, divorced, or separated (81)||32.4 (0-100)||.42|
|Married (124)||35.8 (0-100)|
|High school or less (67)||34.6 (0-80)||.99|
|College or above (138)||34.6 (0-100)|
|Model-generated lifetime BC risk|
|≤10 (38)||23.3 (0-50)||.0008|
|>10 (154)||36.5 (0-100)|
|Yes (154)||34.9 (0-100)||.52|
|No (48)||32.0 (0-80)|
|Patient||Age at diagnosis, y|
|20-49 (112)||37.8 (0-100)||.09|
|50-80 (90)||31.2 (0-100)|
|I (43)||31.9 (0-70)||.80|
|II (82)||34.0 (0-100)|
|III (50)||31.9 (0-80)|
|IV (11)||40.0 (20-50)|
|Time since BC diagnosis|
|<12 mo (146)||36.2 (0-100)||.32|
|12-24 mo (33)||29.1 (0-70)|
|>24 mo (18)||30.0 (10-50)|
The lifetime breast cancer risk for each sister was estimated with the Tyrer-Cuzick model using the IBIS program. All 205 pedigrees, as reported by the breast cancer patients, were entered into the IBIS program. The mean model-generated lifetime risk of breast cancer was 16.6% (range, 3%-61.9%). The risk estimates generated by the model were then compared with the sisters' current subjective risk estimates. There was no significant correlation between the women's current perceived lifetime risk of breast cancer and the risk estimates generated by the model (correlation coefficient [r], 0.14; P = .13). The majority of women (77.2%) overestimated their risk of developing breast cancer. The average degree of overestimation was 2.08 times. The degree of overestimation of one's lifetime risk of breast cancer was found to be correlated with the total level of cancer-related distress (Pearson r, 0.25; P = .005). A sister's current perceived lifetime risk of breast cancer was correlated with her total level of cancer-related distress (Pearson r, 0.30; P = .0006).
The mean level of total cancer-related distress (as measured by the IES) was 25.6 (range, 0-73). The mean score for the intrusion scale was 14.5 (range, 0-35) and the mean score for the avoidance scale was 11.1 (range, 0-38). Thirty-four women scored in the subclinical range (0-8), 67 women scored in the mild range (9-25), 79 women scored in the moderate range (26-43), and 25 women scored in the severe range (≥ 44). A total of 103 women (50.4%) scored in the moderate or severe distress range. In the univariate analysis (Table 3), sister factors (including age, marital status, current perceived breast cancer risk, and model-generated breast cancer risk) and index breast cancer patient factors (including age at diagnosis and stage of breast cancer) were found to be predictive of cancer-related distress. In the multivariate analysis (Table 4), the only significant predictor of cancer-related distress was current perceived lifetime risk of breast cancer (P = .04), with sisters with a perceived lifetime risk of breast cancer of > 20% being more than twice as likely to experience high levels of cancer-related distress compared with women with a perceived lifetime risk < 20%.
|Factors||Variables (Frequencies) N=205||Mean Total IES (Range) N=205||Pa|
|20-49 (92)||30.0 (0-69)||.0002|
|50-80 (113)||22.1 (0-73)|
|Single, divorced, or separated (81)||23.1 (0-55)||.05|
|Married (124)||27.3 (0-73)|
|High school or less (67)||28.0 (0-69)||.12|
|College or above (138)||24.3 (0-73)|
|Perceived lifetime BC risk|
|≤20 (52)||18.4 (0-50)||<.0001|
|>20 (78)||30.4 (3-73)|
|Model-generated lifetime BC risk|
|≤10 (38)||17.9 (0-73)||.0002|
|>10 (154)||28.0 (0-69)|
|Yes (154)||25.1 (0-73)||.49|
|No (48)||26.8 (0-55)|
|Patient||Age at diagnosis, y|
|20-49 (112)||29.1 (0-69)||.0003|
|50-80 (90)||21.3 (0-73)|
|I (43)||21.4 (0-55)||.04|
|II (82)||26.3 (0-69)|
|III (50)||24.8 (0-59)|
|IV (11)||35.6 (9-73)|
|Time since BC diagnosis|
|<12 mo (146)||26.3 (0-73)||.48|
|12-24 mo (33)||23.9 (0-55)|
|>24 mo (18)||22.4 (0-55)|
|Factor||Variables (Frequencies) N=205||Univariate ORa (95% CI)||P||Multivariate ORb (95% CI)||P|
|41-60 (129)||3.59 (1.17-7.57)||.003||1.27 (0.33-4.83)||.83|
|≤40 (31)||4.35 (1.64-11.6)||1.20 (0.24-5.87)|
|Single, divorced, or separated (81)||1||0.14|
|Married (124)||1.52 (0.86-2.67)|
|High school or less (67)||1||.37|
|College or above (138)||0.77(0.43-1.38)|
|Model-generated lifetime BC risk|
|>10 (154)||4.41 (1.96-9.95)||2.18 (0.54-8.82)|
|Yes (154)||0.64 (0.33-1.24)|
|Perceived lifetime BC risk|
|>20 (78)||3.02 (1.46-6.28)||2.35 (1.04-5.30)|
|Patient||Age at diagnosis, y|
|20-49 (112)||2.54 (1.44-4.50)||1.91(0.91-3.99)|
|II (82)||1.53 (0.72-3.24)||.17|
|III (50)||1.80 (0.79-4.10)||.16|
|IV (11)||2.68 (0.68-10.6)|
|Time since BC diagnosis|
|<12 mo (146)||1||.12||1||.14|
|12-24 mo (33)||0.54 (0.25-1.17)||.21||0.53 (0.23-1.23)||.59|
|>24 mo (18)||0.53 (0.20-1.44)||0.74 (0.24-2.25)|
In the current study, we report that sisters of patients newly diagnosed with breast cancer experience high levels of cancer-related distress and have elevated perceptions of personal breast cancer risk. All of the sisters included in the current study had no other family history of breast cancer other than the sister who was recently diagnosed. This was their first experience with a first-degree or second-degree relative being diagnosed with and treated for invasive breast cancer.
In the current study, more than one-half of the sisters scored in the moderate-to-severe distress ranges as measured by the IES. The mean levels of cancer-related distress in this group of sisters of patients newly diagnosed with breast cancer can be compared with other populations. We have previously reported on unselected Jewish women who received positive BRCA1 (breast cancer 1, early onset) and BRCA2 (breast cancer type 2 susceptibility protein) genetic test results and, as a result, had a lifetime risk of breast cancer that approached 80%.18 In the current study, the lifetime risk of breast cancer in the sisters was estimated to be 14.5%. Although there was an extreme difference in lifetime risks of breast cancer between the 2 groups, the mean levels of cancer-related distress were almost identical (25.3 in the Jewish BRCA mutation carriers and 25.6 in the current study). In other studies of BRCA1 and BRCA2 mutation carriers, the levels of distress reported have generally been lower. Schwartz et al reported on the 6-month psychological outcomes of 93 relatives undergoing genetic testing for BRCA1 and BRCA2.19 The relatives without a cancer diagnosis with a positive genetic test result had a mean score of 15.
Cancer-related distress has also been measured in women with breast cancer. Koopman et al reported a mean total IES score of 24.2 in women at the time of a breast cancer diagnosis.20 In a Danish study of 3343 women at 3 months after a breast cancer diagnosis, the mean total IES score was 20.1, and decreased to 16.2 at 15 months after diagnosis.21 Changes in distress scores over time have been demonstrated in other studies, and although we did not observe a significant lowering of total IES scores by time since the sister's breast cancer diagnosis, there were few women in the current study who had a mean time since their sister's diagnosis of > 12 months. However, it may be that the year after a sister's breast cancer diagnosis is the most distressing and relatives may require psychosocial support during this time.
Although there has been a considerable amount of research conducted examining the effects of a breast cancer diagnosis on relatives (mostly daughters and spouses), a study by van Dooren et al reported that having a sister with breast cancer was more distressing than having experienced breast cancer in the mother.22 The study included 105 sisters of women with breast cancer, with the majority (71%) diagnosed > 3 years previously, and 42% having died of the disease. This suggests that although distress is high in the first year, as we have described, it remains a distressing experience for an extended period of time.
In the current study, the only significant predictor of cancer-related distress in a sister of a newly diagnosed patient with breast cancer was their perceived lifetime risk of breast cancer. Women with a perceived lifetime risk of breast cancer >20% were more than twice as likely to experience moderate or severe levels of cancer-related distress compared with those with a perceived lifetime risk of breast cancer of < 20%. Interestingly, the index breast cancer factors, including age at breast cancer diagnosis, stage of breast cancer, and time since breast cancer diagnosis, were not found to be predictors of distress in the sister. This suggests that it is the perceived personal risk of breast cancer driving distress, rather than the prognosis related to the index breast cancer.
It has been shown in high-risk women that long-term distress is significantly predicted by baseline levels of distress.23 The authors argued that this finding may reflect that specific personal characteristics are responsible for increased levels of distress and underlies the importance of identifying vulnerable women at an early stage. Cancer-related distress appears to play an important role in the development and maintenance of anxiety and depression in those women at an increased risk of developing breast cancer.9, 10 Greater than one-half of the women in the current study had moderate or severe levels of distress at the time of a sister's breast cancer diagnosis. To decrease the risk of long-term distress in these women, future interventions should be considered.
Many women overestimate their breast cancer risk. These include women with and without a family history of breast cancer.6, 24-29 Breast cancer risk perception in women at high risk of developing the disease is accurately reported by 9% to 57% of women.6, 24-29 In the current study, the women overestimated their lifetime risk of breast cancer by 2.4 times on average. This overestimation was found to be highly correlated with the total level of cancer-related distress. This suggests that future interventions should be targeted toward helping women to have a more accurate perception of their cancer risk, which may lead to lower levels of cancer-related distress.
There are limitations to the current study. Subjects were recruited from 2 academic health centers, which may not be representative of the population studied. In addition, although we attempted to contact all sisters within 1 year of a breast cancer diagnosis in the index case, 25% of the sisters were enrolled after 1 year. However, in our analyses, time since breast cancer diagnosis was not found to be predictive of cancer-related distress in the sister.
The results of the current study suggest that cancer-related distress is high in sisters of newly diagnosed patients with breast cancer in whom there is no other family history of the disease. Specifically, women with a perceived lifetime risk of breast cancer of > 20% experienced the highest levels of distress. Future interventions that target this group should be considered.
Supported by the Canadian Breast Cancer Foundation (Ontario Chapter).
CONFLICT OF INTEREST DISCLOSURES
Dr. Metcalfe is supported by the Canadian Institutes of Health Research and the Ontario Women's Health Council.
- 1Canadian Cancer Society's Steering Committee on Cancer Statistics. Canadian Cancer Statistics 2012. Toronto, ON: Canadian Cancer Society; 2012.
- 3Increased emotional distress in daughters of breast cancer patients is associated with decreased natural cytotoxic activity, elevated levels of stress hormones and decreased secretion of Th1 cytokines. Int J Cancer. 2002; 100: 347–354., , , , , .
- 11Younger women at increased risk for breast cancer: perceived risk, psychological well-being, and surveillance behavior. J Natl Cancer Inst Monogr. 1994;( 16): 171–176., , .
- 23Long-term psychological distress in women at risk for hereditary breast cancer adhering to regular surveillance: a risk profile [published online ahead of print February 7, 2012]. Psychooncology. doi: 10.1002/pon.3039., , , et al.