Introduction

Breast cancer is the most common cancer in women worldwide (and is the second most common cancer overall, after lung cancer), with nearly 1.7 million new cases diagnosed around the world in 2012 [1]. It is currently estimated that there are more than 3.5 million breast cancer survivors in the USA alone [2]. Further, with increased early detection and improved treatment, the number of breast cancer survivors continues to increase [2].

As a result of both diagnosis and treatment, sexual problems are common and distressing among breast cancer survivors and thus are important quality of life issues [3,4,5,6,7,8]. The most frequently reported sexual problems include diminished sexual interest [9, 10] or desire [3, 6, 11], decreased arousal and lubrication [3, 6, 10], and painful intercourse [3, 12] or vaginal penetration [11]. Survivors have been found to have a greater prevalence and persistence of sexual problems than women of the same age without a history of cancer [3, 10, 11, 13]. Previous research has shown that younger age [14], past chemotherapy use [6, 9, 14,15,16], and vaginal dryness [13, 14, 16] are associated with greater sexual problems among breast cancer survivors. Body image disturbance [10, 15, 16] and depression [17, 18] have also been related to more sexual problems.

The vast majority of research in this area has focused on younger women (i.e., those aged 50 years or younger) [9, 10, 12, 15, 18], newly diagnosed breast cancer survivors within 1 year of diagnosis [e.g., 3, 10, 13, 15], or only women who are sexually active [10, 13]. There have been few longitudinal studies and even fewer with samples of more than 200 survivors [e.g., 9, 12, 19].

The present exploratory study is based on a large sample of married/partnered breast cancer survivors across a wide age range. The goal of this research is to gain a better understanding of factors that impact being sexual active and sexual problems among recently diagnosed breast cancer survivors. The primary objectives are (1) to estimate percentages of partnered women who are sexually inactive over the first 2 years post-breast cancer diagnosis; (2) to identify factors related to being sexually inactive; and (3) to evaluate separately, among both the sexually active and inactive survivors, the relationship between sexual problems and cancer treatment-related variables, symptoms, and psychosocial factors.

Methods

Participants and procedure

The data reported in these analyses are based on secondary analyses derived from a prospective, observational study conducted among women who were newly diagnosed (within 8 months) with stage I, II, or III breast cancer [20, 21]. Study eligibility criteria included first-time breast cancer diagnosis, cancer stages I–III, at least 18 years of age, and the ability to read and understand English. Recruitment was conducted at Memorial Sloan Kettering Cancer Center and the University of Texas–Southwestern Center for Breast Care from 2002 to 2006 with follow-up assessments from 2002 to 2008. Women were recruited through hospital clinics and advertisements and were screened for eligibility. A total of 740 surveys were mailed to women deemed eligible from chart reviews or telephone screening; 653 women completed the baseline survey, for an initial response rate of 88%. The Institutional Review Boards at each of the participating institutions approved this study. Consent was implied for participants who returned questionnaires. All study protocols conformed to the Declaration of Helsinki.

Eligible women were mailed a baseline questionnaire to complete and return to the Coordinating Center at Wake Forest School of Medicine. Baseline questionnaires (T0) were completed within 8 months of diagnosis. Follow-up surveys were completed 6 (T1), 12 (T2), and 18 (T3) months following completion of the baseline survey. Retention rate was high with 86.5% of the sample (n = 565) remaining in the study at T3. Only women who reported that they were married or “living in a marriage-type relationship” at T0 and who remained in such a relationship throughout follow-up were included in these analyses. Data on cancer surgeries and treatments were obtained from medical charts completed following primary treatment.

Primary outcomes

Being sexually active

Women were considered to be sexually active if they responded “yes” to a question “Have you been sexually active in the past month?” This question was asked at each time point.

Sexual problems

Sexual problems were assessed with the sexual problem domain of the Quality of Life in Adult Cancer Survivors (QLACS) scale [7, 22]. The QLACS is a quality-of-life measure containing 47 items and 12 domains specifically designed for cancer survivors. The QLACS has previously been shown to have excellent psychometric properties for both short- and long-term cancer survivors [7, 22]. The sexual problem domain consists of four items: “You were bothered by being unable to function sexually,” “You were dissatisfied with your sex life,” “You lacked interest in sex,” and “You avoided sexual activity.” Items are scored on a 1 (never) to 7 (always) scale, where a higher score indicates more problems. A total domain score is calculated by summing each item for a range of 4 to 28 (coefficient alpha, 0.83). The QLACS was added to the protocol at T2 and T3.

Independent variables

Independent variables in the analysis included sociodemographics, cancer-related variables, symptoms, and psychosocial variables.

Sociodemographic variables included age at diagnosis (continuous, in years), race (non-Hispanic white/other), college (yes/no), and financial strain based on the question: “how difficult is it to pay for basics like food, housing, medical care, and heating?” (very or somewhat hard versus not hard), and children under the age of 18 living at home (yes/no). These sociodemographic variables were assessed at baseline and were not time-varying.

Cancer-related variables included time since cancer diagnosis, cancer stage at diagnosis (I versus II/ III), chemotherapy (coded as none/current/past in repeated-measures analyses using all survey time points, and as ever/never in analyses at T2 only), radiation therapy (treated the same as chemotherapy), hormonal treatment (coded as current/not current, and so updated in repeated-measures analyses), and surgery type (lumpectomy only, mastectomy with no reconstruction, mastectomy with reconstruction).

Symptoms previously found to be associated with sexual problems were included in analyses: severity of vasomotor symptoms and vaginal dryness (both coded on a four-point (0–3) ordinal scale corresponding to “none” to “severe”), as used in the Women’s Health Initiative [23]. Symptom scores were assessed at each survey and thus were time-varying in repeated-measures analyses.

Psychosocial variables included depressive symptoms, as assessed by the Beck Depression Inventory (BDI) [24], and perceived attractiveness, as assessed by the Lasry body image scale [25]. Also included was the Illness Intrusiveness Scale that assesses the degree to which breast cancer diagnosis and treatment interfered with 13 life areas: health, diet, paid work, active recreation, passive recreation, financial situation, relationship with spouse, sex life, family relations, other social relations, self-expression, religious expression, and community [26]. Three items were added to the standard scale that especially impact younger women: family responsibilities, social activities, and work around the house. For each item, respondents rated the degree that their illness or its treatment interfered with that area, based on a seven-point scale, ranging from 1 (not very much) to 7 (very much). The overall illness intrusiveness score is the sum of the 16 items (Cronbach’s alpha, 0.93). The illness intrusiveness sex life question was removed for the analyses where illness intrusiveness was a predictor in regression models of the sexual problem score. All psychosocial variables were assessed at each survey and thus were time-varying in repeated-measures analyses.

We also included baseline menopausal status (pre- or peri-menopausal versus post-menopausal).

Statistical analyses

To describe the sample at baseline, we computed means and standard errors for continuous variables and proportions for categorical variables. For our first objective, we used data from all survey time points to evaluate associations with sexual activity status (active versus inactive) throughout follow-up. For our second objective, which focused on sexual problems of cancer survivors, we used data on problems reported on the T2 survey, when the women were 12–20 months post-diagnosis. We selected this time period to allow us to focus on the period immediately after completion of primary treatment. Our analyses of sexual problems at T2 were conducted separately among the sexually active and the sexually inactive.

For our first objective, we modeled the association between the dichotomous outcome of sexually active versus inactive at all four survey time points using a repeated-measures logistic regression model that predicted sexually active versus inactive status at each time point. Independent variables in this model included age at diagnosis, race, presence of children under 18 years of age in the home, difficulty paying for basics, post-menopausal status, stage at diagnosis, surgery type, chemotherapy status, radiation therapy status, hormone therapy status, vasomotor symptoms and vaginal dryness, depressive symptoms, perceived attractiveness, and illness intrusiveness.

To address the second objective related to sexual problems, we focused on the QLACS sexual problem scores from the T2 interview, which was completed between 12 and 20 months post-diagnosis. We first examined differences by sexual activity status in mean scores on the various QLACS sexual problem items and the overall sexual problem score, using independent sample t tests. We then examined multivariable associations with the overall sexual problem domain score, stratifying these analyses by sexual activity status (active and inactive) at T2. In the multivariable linear regression models, we used the same list of predictor variables as listed above, though only baseline levels of these variables were used. We also included time since diagnosis (in months).

All analyses except the repeated-measures logistic model (done in PROC GENMOD) were complete-case analyses; GENMOD assumes that data are missing at random and utilize full information maximum-likelihood methods with respect to missing data.

We used a two-tailed alpha level of 0.05 throughout. All analyses were conducted using SAS 9.4 (SAS Institute, Inc., Cary, NC).

Results

Sample characteristics

The analytic sample for the present analyses consists of 457 breast cancer survivors who reported throughout the study that they were married (n = 438; Table 1) or “living in a marriage type relationship” (n = 19). We restricted our sample to married/partnered women because they have an “opportunity” for sexual activity, and lack of a sexual partner is a primary reason women report for being sexually inactive use [14, 27]. The women in our sample were predominantly white (94%), reported little difficulty paying for basics (89%), and were predominantly college-educated (65%). Just over half (53.4%) were diagnosed with stage I disease, and in terms of type of surgery, the majority (64.6%) had a lumpectomy only. Sixty-nine percent reported receiving chemotherapy (69%). The mean age at diagnosis was 54.4 years.

Table 1 Sample characteristics of women who were married or partnered throughout follow-up (n = 457)
Full size table

Being sexually active

At T0, 52.4% of women reported being sexually active in the past month. This percentage increased to 59.3% at T1, and then increased only slightly after that, to 59.8% at T2, and 60.7% at T3 (Table 2). The sample size declined modestly over time, and a small number of women at each time point did not answer this question. There was considerable variability in sexual activity status by age, with women under age 45 (75% at T0 to 80% at T3) and women aged 45–54 (60% at T0 to 66% at T3) being more likely to be sexually active than were women over age 64. However, women aged 55–64 years showed the greatest increase over time; only 41.8% were sexually active at T0, but 60.4% reported being sexually active by T3.

Table 2 Partnered women who report being sexually active in the past month by age group and time point
Full size table

In a multivariable repeated-measures logistic regression analysis on the dichotomous outcome of sexually active (yes/no) assessed at each of the four time points, the following variables were significantly predictive of inactivity: older age (p < 0.001), past chemotherapy (compared to none; p = 0.03), depressive symptoms (p < 0.0001), and lower perceived attractiveness (p = 0.002) (Table 3).

Table 3 Results of the multivariable logistic regression analysis of sexually active versus non-active status
Full size table

Sexual problems

Table 4 shows the mean scores on the QLACS sexual problem items and total sexual problem domain score at T2 by sexually active/inactive status at T2. Women who were not active reported significantly more problems than did those who were active for all areas except for being unable to function sexually. For both groups, a lack of sexual interest had the highest problem rating. The total problem domain score was significantly higher for inactive women (14.1) than for sexually active women (10.1). To place these domain scores in context, it is worth pointing out that among the sexually active women, the sexual problem domain was rated more problematic than some of the other problem domains of the QLACS: negative feelings (9.8), cognitive problems (9.6), and pain (8.4). Among the inactive women, the sexual problem domain score was higher than the cognitive problems (10.3), pain (9.4), fatigue (12.7), and negative feelings (10.5) domains. Out of 47 individual items on the QLACS, the “lack interest in sex” item scored as the fourth most problematic area for the sexually active women and was the most problematic item among the inactive women.

Table 4 Mean responses on the sexual problem domain of the QLACS at T2 by sexually active versus non-active status
Full size table

In stratified multivariable linear modeling analyses with sexual problem domain score at T2 as the outcome and containing all covariates, depressive symptoms were significantly related to greater sexual problems among both sexually active (p < 0.001) and inactive (p = 0.002) women, as was vaginal dryness (p < 0.001 for both groups; Table 5). Among the sexually active women, younger age at diagnosis (p = 0.02), less illness intrusiveness (p = 0.04), and lower perceived attractiveness (p = 0.009) were related to more problems. Among those who were not sexually active, being on hormone therapy was significantly associated (p = 0.03) with fewer predicted sexual problems. Hormone therapy was the only cancer treatment-related variable that was significantly related to sexual problem score for either group.

Table 5 Results of the multivariable regression analysis of the QLACS sexual problem domain score at T2, stratified by sexually active versus non-active status
Full size table

Discussion

This is perhaps the largest longitudinal study to date that has examined sexual activity and sexual functioning problems among breast cancer survivors over a 2-year period following diagnosis. Among married/partnered women, 52% reported being sexually active in the past month within the first 8 months of diagnosis. This percentage increased to 59% at T1 and 61% at T3, which was 18–24 months post-diagnosis.

Although our study did not have a comparison group of women without cancer, several other studies provide indirect comparative data, suggesting that a lower percentage of breast cancer survivors are sexually active 2 years post-diagnosis compared to women without cancer. Dorval and colleagues reported that 83.3% of partnered women without cancer reported being sexually active in the past year [28]. Another study found that 71% of women aged 40–69 who were in a significant relationship reported being sexual active in the past year [29]. Although we found that 29% of sexually active women and 62% of inactive women reported that lack of interest occurred more than sometimes, Broekel and colleagues found that only 13% of women without cancer reported that lack of sexual interest was a least somewhat of a problem [30].

In addition to being older, inactive women reported more depressive symptoms and lower perceived attractiveness and were less likely to have had chemotherapy. They also reported that they had less interest in sex and were more likely to avoid sexual activity than women who were sexually active. Although inactive women were not asked the extent to which they wanted to be sexually active, this group reported at least sometimes being dissatisfied with their sex life. Among the items on the illness intrusiveness measure, sex life had the highest mean score out of 16 life areas for the active and inactive women under age 54 (the mean was 5.6 on a seven-point scale for the inactive women under age 45 and 4.1 for those age 45–54). These findings suggest that sexual functioning is a potential concern among partnered breast cancer survivors who have not been sexually active and that studies of sexually active women only are missing this important group of women.

For both active and inactive women, vaginal dryness and depressive symptoms were related to greater sexual functioning problems. Similar findings have been observed among other studies of breast cancer survivors [13, 14, 16,17,18], as well as in women without cancer [31], but these factors pose a particular problem for breast cancer survivors [32]. Although estrogen therapy can help reduce vaginal dryness, this treatment is generally contraindicated for cancer survivors [33]. Breast cancer survivors are also at higher risk for depressive symptoms [32].

In addition to the above factors, among those survivors who were sexually active, we also found that younger age and lower perceived attractiveness were related to more problems, findings that are consistent with other studies [9, 10, 15, 16]. Psychosocial counseling and/or educational programs, such as the American Cancer Society sponsored program “Look Good Feel Better,” might help women with body image concerns. One recent study found that Internet-based cognitive behavioral therapy improved body image and sexual functioning among breast cancer survivors aged 18–65 [34].

Implications for practice

Responses to the QLACS sexual problem measure and the relatively high ranking of sex life on the illness intrusiveness measure, particularly for women aged 54 and under, emphasize the importance of sexual functioning relative to other life areas, among many breast cancer survivors. Previous studies, however, have found that sexual functioning often is not discussed between patients and providers [8, 35]. One study found that 70% of breast and gynecological cancer patients expressed concerns about sexual function and 66% had unmet needs for consultation with a sexual health expert [35]. Recent guidelines from the American Society of Clinical Oncology (ASCO) recommend that health care team members should initiate a discussion with all cancer patients regarding sexual health and dysfunction resulting from cancer or its treatment [36]. These guidelines further recommend that psychosocial and/or psychosexual counseling be offered to all patients over age 18 with cancer. Clinician decisions about broaching the topic of sexual health can sometimes be based on erroneous assumptions about the relevance of sexual health concerns [5, 37].

Clinicians should also recognize that the relationship between sexual inactivity/problems and depression can be complicated and is likely reciprocal. A diagnosis of breast cancer and its subsequent treatment can lead to depressive symptoms which can contribute to the avoidance of sexual activity [17, 18]. On the other hand, cancer treatments can also have a direct impact on vaginal dryness and painful sex leading to sexual problems, which may in turn lead to depressive symptoms [4, 38]. To address these problems, products such as vaginal rings and moisturizers are available to help alleviate vaginal dryness which may in turn help reduce depression. Behavioral or pharmaceutical treatments of depression may in turn improve sexual functioning. The ASCO guidelines provide a brief summary of the effective treatments for these and other specific types of sexual problems [36].

Limitations

There are several limitations of this study. The sample was relatively homogeneous (mostly white, educated, and partnered women), which limits its generalizability. Although the QLACS items can be answered regardless of sexual orientation (since no male-specific sexual acts are referred to), we do not know if results would differ by sexual orientation, which was not assessed in this study. A few studies have concluded that sexual minority breast cancer survivors and heterosexual survivors are similar in sexual frequency and sexual satisfaction, while lesbian breast cancer survivors report less disruption in their sexual relationships [39, 40] although research on sexual minority women is quite limited [41]. Second, like most studies, we did not have a measure of sexual activity/functioning prior to diagnosis and are unable to examine the impact of cancer diagnosis and treatment on change in sexual activity and functioning. Third, relationships with sexual partners and perceived partner dysfunction can be important factors [16], but they were not explored. Additionally, menopausal status was only obtained at baseline, and previous studies have shown that change in menopausal status can influence sexual functioning [10, 19]. Lastly, the sample data were collected approximately 10 years ago and may not reflect advances in treatment regimens since this time. However, chemotherapy, endocrine therapy, surgeries, and radiation continue to have a large effect on woman’s sexual health and function [38, 42, 43].

Strengths of the analyses include a large sample of breast cancer survivors of all ages, a wide range of potential predictors, and a focus on the newly post-treatment survivorship time period. Our findings suggest that sexual functioning is a concern for many (though not all) women who are partnered but not currently sexually active, a group that is often understudied. Our data suggest that assumptions about the relative importance of sexual health cannot be based only on a woman’s sexual activity status. Sexual inactivity among partnered women has important implications for intimate relationships. As sexual activity stops, other forms of intimacy and affection such as touching and physical contact may also decline [4, 43]. Both survivors and partners have reported that these changes can lead to isolation and sadness, and are often accompanied by poor communication [44]. Although other quality of life issues may improve after primary treatment, sexual health concerns are among the complications that can worsen over time [33, 45]. Moving forward, providers from a variety of clinical specialties need to be prepared to address sexual health discussions and offer appropriate resources and referrals.