Introduction

Following breast and prostate cancers, colorectal cancers have the highest likelihood of long-term survival [1]. As cancer survivors live longer, it becomes increasingly important to understand the factors affecting their quality of life (QoL) including their sexual function and related outcomes [2]. Sexual difficulties are common for men and women with colorectal cancer [3], affecting as many as many as 88 % of men [4] and around half of women [3]. These difficulties have been shown to persist for as long as 10 years after treatment completion [5] and are associated with worse overall QoL and disease impact [6, 7]. A potential contributor to sexual difficulties for colorectal cancer patients is the presence of a gastrointestinal ostomy [3, 8]. Ostomy surgery can lead to sexual dysfunctions for men and women such as erectile dysfunction and dyspareunia (pain during sexual intercourse), respectively [9, 10]. Furthermore, patients with ostomies report disturbances in body image [11, 12] and logistical obstacles to sexual activity such as unpleasant odors during sex [13], among other issues.

Despite the strong evidence showing ostomy-related sexual difficulties, little is known about whether sexual difficulties persist after ostomies are reversed [14, 15]. Establishing whether the effects of ostomy on sexual outcomes and body image endure after an ostomy is reversed is important given that temporary stomas are increasingly being performed for patients with rectal cancer [16]. This information would also lead to developing appropriate interventions for these at-risk groups. In addition, prior research has focused on physiological sexual dysfunctions rather than psychosocial outcomes (e.g., perceived adjustment to sexual changes, body image distress, and depression)—constituting a critical gap in this literature. In particular, few studies have considered the potential contribution of demographic and treatment-related confounding factors in examining sexual outcomes in ostomy patients [14, 17] and none have considered depressive symptoms as a potential confounding factor—which is important as depressive symptoms may be related to both body image distress and sexual dysfunction in ostomy patients [18, 19]. While the evidence for worse depression for colorectal cancer patients with vs. without ostomies is mixed [8, 11, 20, 21], depression appears to be a risk factor for worse sexual function for colorectal cancer patients [22]. Moreover, body image was shown to predict depression [18] and contribute to sexual problems such as low sexual interest [19] for ostomy patients. Finally, most research studies examining sexual function in colorectal cancer patients have examined sexual function as a secondary outcome, leading to inclusion of brief or unstandardized measures of sexual outcomes [3]. Including a range of comprehensive, standardized measures would represent a methodological improvement by providing a more complete understanding of the sexuality of those with ostomies.

Accordingly, our objective for this study was to use comprehensive validated measures to compare colorectal cancer patients by ostomy classification group (never, past, and current) with respect to sexual function, perceived impact of their disease and treatment on sexual function, and body image distress. We hypothesized that (1) the study sample overall would report sexual function below established norms and that (2) both ostomy groups (i.e., past and current) would report worse sexual outcomes and body image distress than the group without an ostomy. These hypotheses were influenced by prior research suggesting that psychological and relational factors (e.g., getting “out of the habit” of intimacy) are related to sexual difficulties for those with either temporary or permanent ostomies [13].

Methods

Patients

Men and women older than age 21 with a diagnosis of colorectal cancer were eligible for this research. Participants were recruited between December, 2009 and April, 2012 from Johns Hopkins Sidney Kimmel Comprehensive Cancer Center. Both partnered (married or cohabiting) and unpartnered patients completed sexuality items in all study analyses. Individuals with colostomies or ileostomies were included (see Table 1).

Table 1 Demographic and medical characteristics for the total sample and by ostomy group
Full size table

Procedure

Participants in the current study were involved in a larger prospective study examining physical and emotional predictors of sexual QoL in colorectal cancer. Baseline data focusing on patients’ sexual outcomes and body image distress are presented here. Candidates were recruited directly in the clinic or through mailings to Johns Hopkins tumor registry patients (diagnosed within past 5 years). Two hundred fifty-eight surveys assessing patients’ emotional, physical, and sexual health were administered and 143 (55 %) were returned. Two patients who completed the survey were excluded because they did not have colorectal cancer. The final sample included 141 individuals with colorectal cancer; 57 % were recruited in the clinic (n = 80); the rest (n = 61) were recruited through mailings. Patients recruited in the clinic were younger and were more likely to have metastatic disease, be receiving treatment, have received chemotherapy, and have a shorter length of time since diagnosis than patients recruited through mailings (p values < .01). Patients recruited in the clinic reported greater depressive symptoms (M = 7.5; SD = 5.94) than those recruited through mailings (M = 5.46; SD = 5.79; p = .04); no other differences were found. Institutional Review Board approval was obtained, and patients provided informed consent through completion of the baseline survey. All surveys were returned through the mail, and patients were mailed parking coupons for completing study surveys.

Measures

Medical impact on sexual function

The Sexual Function Questionnaire (SFQ) [23] was developed specifically for use in cancer populations [23, 24] while the Medical Impact subscale assesses the impact of a medical condition or treatment on patients’ sexual lives and sexual function. The Medical Impact subscale consists of five questions assessing the impact of the patient’s medical condition (in this case, “colorectal cancer or its treatment”) on aspects of his or her sex life, resulting in a mean score. Items assess the impact on desire, arousal, orgasm, overall impact, and degree of adjustment to sexual difficulties. Mean scores were calculated when ≥80 % of items (4/5) were completed. Higher scores indicate greater impact.

Female sexual function

The Female Sexual Function Index (FSFI) [25] is a 19-item questionnaire that assesses female sexual function and has been used in many cancer studies [26–28]. The scale provides scores on six domains: desire, arousal, lubrication, orgasm, satisfaction, pain, and a total score which we report on in the current study to allow for comparisons with established cut-off scores. Respondents report on the past 30 days. Higher scores indicate better function.

Male sexual function

The International Index of Erectile Function (IIEF) [29] is a 15-item widely used multidimensional scale assessing sexual functioning in men [28, 30, 31]. The IIEF consists of five domains: erectile function, orgasmic function, sexual desire, intercourse satisfaction, overall satisfaction, and a total sexual function score. In the current study, we report on both total sexual function score to provide data on overall sexual function across a range of domains and on the erectile function domain score to allow for comparisons with established clinical cut-off scores on this domain. Respondents report on the past 30 days. Higher scores indicate better function.

Body image distress

Body image distress was assessed using the Body Image Scale (BIS) [32], a ten-item scale developed for use in cancer patients that assesses body image changes and distress due to cancer and its treatment. Higher scores indicate greater symptoms or distress.

Depressive symptoms

The Center for Epidemiologic Studies Depression Scale—Short Form is a ten-item, self-report scale designed to measure symptoms of depression in the general population [33]. Given the potential importance of depression as impacting sexual function and body image distress, as well as inconsistent prior findings pertaining to differences in depressive symptoms for those with and without ostomies [11, 12, 20, 34], we examined group differences in this outcome and considered depressive symptoms as a potential covariate in comparisons on other sexual outcomes. We did not hypothesize as to which groups would differ significantly on this outcome.

Medical information

Information on medical characteristics including on tumor site (colon/rectum), disease stage, treatment status (on treatment or completed treatment), types of treatments and surgery, ostomy status and type, and length of time since diagnosis were obtained through self-report and/or medical chart review.

Statistical methods

Ostomy groups (current, past, and none) were compared on demographic and health status variables, including depressive symptoms, using Chi-square or Fisher’s exact tests for categorical variables and general linear models for continuous or semi-continuous variables. Next, descriptive analyses were conducted on female and male sexual function to facilitate comparisons with established cut-off scores. General linear models were used to test for ostomy group differences in the following outcomes in unadjusted models: medical impact on sexual function, female and male sexual function, and body image distress. Then, the following covariates were included in initial adjusted models: age, gender, pelvic surgery, metastatic disease, chemotherapy, radiation therapy, and depressive symptoms. Our selection of covariates was motivated by both sample characteristics and associations reported in prior literature [8, 11, 14, 28, 35]. Nonsignificant covariates (p > .10) were removed from the final models (see Table 2). Surgery was classified as either pelvic (lower anterior resection and abdominoperineal resection/proctectomy) or nonpelvic (colectomy or hemicolectomy). Patients who received other surgeries (transanal excision, n = 3; exploratory laparotomy, n = 1; sigmoidectomy, n = 1), those who had not received surgery at the time of the baseline survey (n = 5), and those who had missing data for type of surgery (n = 7) were not included in analyses adjusting for pelvic surgery. Considering the high number of comparisons, pair-wise differences were examined and multiple comparisons corrected by the Bonferroni procedure within each analysis. To maximize the use of available data, means of existing items were imputed for missing items when the majority of items on a scale were completed. Total FSFI and IIEF scores were calculated when all domain scores were available from which to calculate a total score. When too few items were completed to calculate a score (e.g., a complete scale was left blank) for a participant, those data were excluded. Analyses were performed using SPSS Version 20 (SPSS, Chicago, IL). All tests were two-tailed, and p values below .05 were considered statistically significant.

Table 2 Differences (means and confidence intervals) on sexual outcomes and body image distress by ostomy group
Full size table

Results

Sample characteristics

Sample characteristics for the total sample and by ostomy group are presented in Table 1; 12.7 % of the sample currently had an ostomy (n = 18), 17.7 % previously had an ostomy (n = 25), and 69.5 % had no history of an ostomy (n = 98). Most participants (73.8 %) had colon cancer and half the sample (49.6 %) had Stage IV disease. Almost all patients had undergone surgery; most had also undergone chemotherapy and/or radiation therapy.

Comparisons among ostomy groups on demographic, medical variables, and depressive symptoms

The ostomy groups did not differ significantly on gender, race, marital status,Footnote 1 highest educational degree obtained, currently receiving treatment, or metastatic disease (p values ≥ .23). The groups differed marginally on age (p = .05); the past ostomy group was slightly older than the current ostomy group (p = .08). The group with no ostomy history was more likely to have colon cancer than the other groups (χ 2 = 28.40, p < .001). Both ostomy groups were significantly more likely to receive pelvic surgery (χ 2 = 44.75, p < .001), radiation therapy (Fisher’s exact = 36.13, p < .001), and post-operative complications (χ 2 = 6.38, p = .04), and marginally more likely to receive systemic chemotherapy (Fisher’s exact = 5.46, p = .06) than those with no ostomy history. On depressive symptoms, no differences were found among those with a past ostomy (M = 5.39; SD = 3.36), current ostomy (M = 8.08; SD = 6.10), or no ostomy history (M = 6.49; SD = 6.25, p = .33).

Comparisons among ostomy groups on sexual outcomes

Medical impact on sexual function

Most participants completed the SFQ (113/141; 80.1 %).Footnote 2 Means and 95 % confidence intervals for the three groups from unadjusted and adjusted analyses are shown in Table 2. In adjusted analyses, estimated marginal means are shown. When controlling for chemotherapy and depressive symptoms, ostomy status distinguished significantly among the groups (p < .001). Both the past and current ostomy groups had worse medical impact on sexual function than those with no ostomy history (p < .001). Age, gender, pelvic surgery, metastatic disease, and radiation therapy were removed from the final model.

Female sexual function

Most women (42/59; 71.2 %) completed the FSFI; of them, most (74 %) reported sexual function below the FSFI clinical cut-off (26.55) indicating sexual dysfunction [36]. All women with a current ostomy scored in the dysfunctional range (10/10), compared with 75 % of women with a past ostomy (3/4) and 64 % percent of women with no ostomy history (18/28). Given the small sample sizes in the past and current ostomy groups, we collapsed data across these two groups and compared this new group (past or current ostomy) with those who never had an ostomy. In adjusted analyses, ostomy status differentiated female sexual function when age and depressive symptoms were covaried (p = .01); the past/current ostomy group had lower sexual function than those who never had an ostomy. Pelvic surgery, metastatic disease, chemotherapy, and radiation therapy were removed from the final model. The two groups did not differ on partnered status (χ 2 = .30, p = .58).

Male sexual function

Most men in the overall sample (76/82; 92.7 %) completed the IIEF; of them, 65 % scored below the IIEF Erectile Dysfunction clinical cut-off of 25 indicating dysfunction [31]. Mean scores for all three groups fell in the moderate range of erectile dysfunction; 67 % of men with a current ostomy scored in the dysfunctional range for erectile dysfunction (8/12), compared with 82 % of men with a past ostomy (9/11) and 60 % of men with no ostomy history (32/53). There were no group differences by ostomy status on male overall sexual function or erectile function (p values ≥ .54). For both overall sexual function and erectile function, age, metastatic disease, and radiation therapy were significant covariates; for overall sexual function, depressive symptoms were also significant.

Body image distress

Complete data were available for the BIS. The three groups differed significantly on body image distress when adjusting for age, gender, and depressive symptoms (p < .001). Current ostomy patients had significantly worse body image distress than those with no ostomy history (p < .001). Pelvic surgery, metastatic disease, and radiation therapy were removed from the final model.

Discussion

Sexual disturbances are common and problematic for many colorectal cancer patients. In this study, patients in both the past and current ostomy groups reported greater perceived impact on sexual function than those who never had an ostomy, and this did not differ by gender. Furthermore, women with either a past or current ostomy (combined into one group) reported worse sexual function than women who never had an ostomy. By contrast, the past ostomy group reported similar body image distress to those who never had an ostomy, suggesting that unlike impact on sexual function, body image may improve with ostomy reversal. Thus, sexual difficulties can persist after ostomy reversal and may be attributable to factors other than body image distress (e.g., difficulties making behavioral adjustments and long-term physical dysfunction).

Importantly, sexual dysfunction was common irrespective of ostomy status and gender. All women with a current ostomy reported sexual function mean scores in the dysfunctional range based on established cut-off scores, and most women in the two other groups also fell in the dysfunctional range. Similarly, most men in the overall study sample reported erectile dysfunction in the dysfunctional range. The lack of group differences on male sexual function contrasts somewhat with prior research which found worse sexual function for male patients with ostomies [8]. This may be attributable to characteristics of the study sample, such as a somewhat lower age across all three groups than in other studies [17]. That group differences were found for some measures and not others highlights the importance of assessing multiple dimensions of sexuality. Taken with prior research [13], findings suggest that colorectal cancer and its treatment are associated with sexual difficulties and some difficulties appear more pronounced for those with ostomies as part of their treatment.

Differences in sexual outcomes by ostomy status could not be attributed to these patients being older, of a particular socio-demographic group, more or less likely to be partnered, or to receiving pelvic (vs. non-pelvic) surgery, chemotherapy, or radiation therapy. Depressive symptoms, which have not been previously analyzed as a covariate in similar studies, contributed variance in final models, but did not significantly differ among ostomy groups. As some current ostomy patients in this study had temporary, rather than permanent ostomies, this may have lowered the potential depression levels of current ostomy patients. It is interesting that the groups in this study differed on body image but not depressive symptoms, even though body image distress and depression have been shown to be associated in prior studies of ostomy patients [18]. Taken with prior research, these findings suggest that ostomy patients with poor body image are at particularly elevated risk for psychological distress.

The current study is strengthened by the inclusion of a range of validated, comprehensive measures and a prospective study design. However, several aspects of the current study limit the ability to generalize findings. First, the data analyzed were cross-sectional and do not support causal interpretations. Future prospective studies examining change in sexual function from pre- to postsurgery are necessary to understand the persistence of poor sexual outcomes and body image distress after ostomy reversal. Second, the two ostomy groups had relatively few subjects. Future studies with larger samples are needed to replicate findings. About a quarter of women did not complete sexual function items, reflecting a challenge in collecting female colorectal cancer patients’ sexual function data seen in prior research [17]. In addition, half of the study sample had advanced disease, making it difficult to generalize to patients with early stage disease. While we adjusted for metastatic disease, our sample size limited the ability to adjust for factors such as length of time from treatment, and future studies should do this. Third, we were not able to compare study completers with those who refused to participate on demographic or clinical factors; future studies should conduct such comparisons. As colon and rectal cancer may have different effects on sexual function [22], future studies should be powered to allow for these two related but different cancers to be examined separately. Finally, as the ostomy patients in this sample also underwent pelvic surgery and often radiation, we cannot state with certainty whether poorer sexual outcomes on some measures are due to the ostomy itself or to the consequences of the pelvic surgery or radiation such as scarring of the vaginal wall and injury to pelvic nerves (hypogastric nerves or nervi erigentes). However, prior research strongly suggests that ostomies affect multiple aspects of patients’ sexuality, including their sexual identities, intimate relationships, and psychological function [13, 35]. Thus, it is doubtful that the associations found in this study between ostomy use and worse sexual outcomes are due exclusively to physical changes resulting from surgery or radiation.

Findings of the study have several clinical implications. First, they suggest that use of an ostomy—even when it has been reversed—is associated with difficulties with sexual adjustment for those with colorectal cancer. Therefore, discussions about the potential impact of colorectal cancer treatment on sexuality may be beneficial. Assessments should be conducted to identify patients with concerns or difficulties, and interventions addressing the particular concerns of ostomy patients could be utilized for patients with the greatest difficulties. The PLISSIT model [37] can guide approaches to managing sexual complaints for ostomy patients [38, 39]. In this model, permission to discuss sexual issues is first given by raising the topic of sexuality, followed by limited information on how the ostomy may affect body image and sexuality, then specific suggestions (e.g., using ostomy covers during sexual activity; emptying the ostomy pouch prior to sexual activity), and finally intensive therapy for patients with severe or complex sexual difficulties. In the context of a busy oncology clinic, an effective strategy may be to have one provider on the treatment team trained to offer brief sexual counseling [40]. An increasing number of comprehensive cancer centers maintain experts on staff able to address the sexual health needs of patients who require intensive assessment and treatment (e.g., psychologists and sex therapists), as we do at our site. When on-site specialists are not available, providers should know of off-site resources and make such referrals when appropriate. Continuing to engage in sexual activity, including non-intercourse physical intimacy, and redefining sexual function and activity are particularly promising strategies that could be incorporated into interventions for those with colorectal cancer [19, 41]. Sexuality is a critical domain of QoL that warrants clinical and research attention for those with colorectal cancer.