Sexual dysfunction in young adult survivors of childhood cancer e A population-based study

Objective: To determine the prevalence of sexual dysfunction and to identify the factors associated with sexual dysfunction in young adult childhood cancer survivors. Methods: All survivors of childhood cancer (aged 19 e 40 years) in Sweden were invited to this population-based study, and 2546 men and women (59%) participated. Sexual function was examined with the PROMIS Sexual Function and Satisfaction Measure. Logistic regression was used to assess the differences between survivors and a general population sample (n Z 819) and to identify the factors associated with sexual dysfunction in survivors. Results: Sexual dysfunction in at least one domain was reported by 57% of female and 35% of male survivors. Among females, dysfunction was most common for Sexual interest (36%), Orgasm e ability (32%) and Vulvar discomfort e labial (19%). Among males, dysfunction was most common for the domains satisfaction with


Introduction
With an increased 5-year survival rate approaching 80%, the population of long-term childhood cancer survivors is growing [1].Only in Europe, there are about 500,000 people living with a childhood cancer experience [2].This is a group known to be at risk of physical and psychological late effects [3,4], and many survivors suffer from chronic health conditions [4].Despite being acknowledged as an integral part of health and quality of life [5,6], data on the effect of childhood cancer on sexual health in young adult survivors are limited.
Sexual dysfunction in childhood cancer survivors can be secondary to physical, psychological and social implications related to the cancer experience [7,8].Abdominal/pelvic or total body irradiation as well as certain types of chemotherapy can result in delayed or arrested puberty, premature menopause and testicular dysfunction [3].Previous studies suggest that both male [9e12] and female [13,14] survivors experience more sexual dysfunction than siblings and peers in similar ages.
Sexual dysfunction has been reported for 20e52% of female and 20e32% of male survivors of childhood cancer [6,13,15].Reported problems include low interest, erectile dysfunction, vaginal dryness, vaginal pain/ discomfort and difficulties enjoying sex [6,13,15e17].Risk factors for sexual dysfunction among survivors include female sex, older age at study, health problems and low income [8].In addition, a negative body image [16e20] and depressive symptoms [9,12,15] have been linked to sexual problems.How treatment and illnessrelated factors affect survivors' sexual function is not clear.Some studies suggest that survivors of certain types of cancer are at higher risk for sexual dysfunction [12e14], but other studies show no such associations [15].
The evidence concerning sexual dysfunction among childhood cancer survivors is hampered by methodological limitations regarding sampling and measures [6,15,21].Large-scale, methodologically rigorous studies using representative samples and validated measures are, thus, needed to gain better knowledge about the extent of sexual problems in survivors and how these are associated with sociodemographic, psychological and clinical factors.Such knowledge can be used to develop targeted screening and interventions for this population.The aim of this study was, therefore, to determine the prevalence of sexual dysfunction and to identify factors associated with sexual dysfunction in young adult childhood cancer survivors.

Design
The study is part of the Fex-Can Childhood project, whose methods are described in a study protocol [22].This nationwide, population-based study combines registry and survey data for cancer survivors and includes a general population sample described in detail elsewhere [23].

Participants
Survivors were identified through the National Quality Registry for Childhood Cancer (NQRCC), which includes all individuals diagnosed with paediatric cancers in Sweden.Eligible participants were diagnosed between ages 0 and 17 and were 19e40 years of age and residents in Sweden at the time of enrolment.Exclusion criteria were as follows: unable to read/write in Swedish, selfreported poor health and/or cognitive impairment that prevented survey completion.
For the comparison group, a random sample (1000 women; 1000 men) from the general population aged 19e40 was identified by the Swedish population registry and invited to the study [23].The same exclusion criteria as for survivors were used.Additionally, individuals who reported treatment for cancer were excluded.

Procedure
The survey was sent to potential participants together with a letter describing the study.Participants could complete the survey via paper or Web, and nonresponders received two reminders.Data for survivors were collected from August 2019 to February 2020 and for the comparison group from April to June 2018.Ethical approval was granted by the Regional Ethical Review Board in Stockholm, Sweden (approval number: 2015/1609-31; 2018/2688-32; 2019/01066; 2019/04603).

Measures
Participants completed established self-report instruments together with study-specific items on use of antidepressants, sexual orientation and sociodemographic factors.
Sexual function was measured by the Patient-Reported Outcomes Measurement Information System (PROMIS) Sexual Function and Satisfaction Measure (SexFS) version 2.0 [24].The following domains were answered by females: Vaginal lubrication, Vaginal discomfort, Vulvar discomfort e clitoral and Vulvar discomfort e labial.The Erectile function domain was answered by males.Four gender-neutral domains were used: Satisfaction with sex life, Interest in sexual activity, Orgasm e ability and Orgasm e pleasure.All respondents completed items related to Interest in sexual activity, whereas remaining domains were completed only by respondents who had been sexually active.In the questionnaire, sexual activity was defined as sex with a partner and/or solo sex including masturbation, oral sex and intercourse.Domain scores were transformed to a T-score metric, where 50 represents the mean for the American general population (standard deviation [SD] Z 10) [24].As recommended by the PROMIS network (http://www.nihpromis.org/), a cut-off of 1 SD from mean was used to define sexual dysfunction.The SexFS has shown adequate construct, content and knowngroups validity and test-retest reliability [24e26].The items were translated into Swedish and linguistically validated in accordance with FACITrans and PROMIS procedures [27].
Body image disturbance was assessed using the Swedish version of the Body Image Scale (BIS).BIS assesses body image discomfort in general (five items) and in relation to cancer and its treatment (five items) [28].In line with previous research on non-cancer populations [29], the comparison group responded only to the five general items.The total sum score was used in this study, with higher scores indicating more disturbance.BIS has shown clinical validity, high test-retest reliability and satisfactory internal consistency [28,30].
Emotional distress was assessed using the Swedish version of The Hospital Anxiety and Depression scale (HADS) [31].HADS has shown satisfactory internal consistency and concurrent validity [32,33].In this study, the two subscales assessing symptoms of anxiety and depression were combined in one overall score (range 0e42), with higher scores indicating greater emotional distress.
Clinical data were retrieved from the NQRCC including cancer diagnosis, date of diagnosis, age at diagnosis and treatment.Diagnoses were classified according to the International Classification of Childhood Cancer, 3rd revision [34].Each child's treatment was categorised by a paediatric oncologist using the Intensity of Treatment Rating scale (ITR-3.0), a psychometrically validated measure of treatment intensity of current treatment protocols in paediatric oncology [35].Different disease and/or treatment modalities were classified according to one of the four intensity levels, from level 1 (minimally intensive) to level 4 (most intensive).

Analyses
Clinical and sociodemographic characteristics of groups were compared by gender using Student's t-tests and chi-square tests.Prevalence of sexual dysfunction was compared between survivors and the comparison group using logistic regression, adjusted for sociodemographics (age at study, education, relationship status, having children and country of birth).To identify the factors associated with dysfunction in survivors, logistic regression models were conducted for each SexFS domain and for dysfunction in two or more domains.The following potential factors were selected a priori based on the literature: age at diagnosis, time since diagnosis, age at study, education, country of birth, relationship status, sexual orientation, use of antidepressants, having children, type of cancer, treatment intensity, body image disturbance and emotional distress.First, each factor was examined in bivariate analyses, using simple logistic regression and chi-square tests as appropriate.Factors associated with dysfunction in the respective domain (P < 0.10) were thereafter analysed using multivariable logistic regression.
All tests were two-tailed with P < 0.05 denoting statistical significance.Analyses were performed using SPSS Statistics version 25 (IBM Corp., Armonk, N.Y., USA).

Study participants
Of the 4328 survivors approached, 90 survivors were excluded due to unknown address (n Z 49), living outside Sweden (n Z 3), cognitive dysfunction (n Z 32) and administrative failure (n Z 6).Study participants were 2546 survivors (59% response rate), of which 1333 females (67% response) and 1213 males (52% response).Table 1 shows the characteristics of the participants.Among survivors with leukaemia (most common diagnosis), 65% of females and 56% of males had received moderately intensive treatment, whereas 35% and 44%, respectively, had undergone very/most intensive treatment.More detailed information on intensity of treatment by diagnosis is provided in Suppl.Table 1.
For survivors, female responders and non-responders did not differ in demographic or clinical characteristics.Among males, responders were older (mean Z 29 and 28 years, respectively; P < 0.05), and time since diagnosis was longer than for non-responders (mean Z 21 and 20 years, respectively: P < 0.05).Sociodemographic characteristics were compared between survivors and the comparison group (n Z 819, 42% response rate).Compared with survivors, more females and males in the comparison group were born outside Sweden, were partnered and had children (Table 1).Moreover, females in the comparison group were older at the time of study.

Prevalence of sexual dysfunction
More than half of female (57%) and one-third of male (35%) survivors reported a dysfunction in at least one domain during the last month, and about one-fifth (22%) of females and one-tenth of males (13%) reported dysfunction in at least two domains (Table 2).Among females, sexual dysfunction was most common in the domains of Interest in sexual activity, Orgasm e ability and Vulvar discomfort e labial.The most commonly reported sexual dysfunction in male survivors concerned Satisfaction with sex life, Interest in sexual activity and Erectile dysfunction.For prevalence of dysfunction by ICCC-3 group diagnosis, see Suppl.Table 2.

Factors associated with sexual dysfunction in survivors
Female survivors who were partnered and with a higher education were less likely to report dysfunction in certain domains (Table 3).Females with children and who were born outside Sweden were more likely to report dysfunction regarding Interest in sexual activity, whereas females who reported a sexual orientation other than heterosexual were less likely to report dysfunction in this domain.Females who had received more intensive treatment and with greater body image disturbance were more likely to report dysfunction in two or more domains.Furthermore, female survivors with greater emotional distress were more likely to report dysfunction in all outcomes.
Males in a relationship were less likely to report dysfunction in two or more domains, and males with a more intensive treatment were more likely to report dysfunction related to Orgasm e pleasure (Table 4).Male survivors with a higher educational attainment were less likely to report dysfunction in the domain Interest in sexual activity.Moreover, male survivors with greater emotional distress and greater body image disturbance were more likely to report dysfunction in two or more domains.

Discussion
This is one of the most comprehensive studies to date on sexual dysfunction in young adult survivors of childhood cancer.About half of the female and onethird of the male survivors reported at least one current sexual dysfunction.Among female survivors, dysfunction was most common in the domains Sexual interest, c According to the Intensity of Treatment Rating (ITR-3.0)[35].
Orgasm e ability and Vulvar discomfort.Among male survivors, sexual dysfunction was most frequently reported in the domains Satisfaction with sex life, Interest in sexual activity and Erectile function.While the prevalence of sexual dysfunction among female survivors was similar to that of women in the comparison group, male survivors reported a higher prevalence of dysfunction for two domains compared to men in the general population sample.Corresponding to what has been reported for the comparison group [23], increased likelihood of sexual dysfunction in one or more domains in survivors was associated with not having a partner, low education, as well as concurrent emotional distress and body image disturbance.Our results furthermore show an increased likelihood of sexual dysfunction in survivors subjected to a more intensive treatment.
In comparison with our general population sample, both female and male survivors were significantly less likely to have a partner and to have children.When controlling for sociodemographic variables, we found few differences in prevalence of sexual dysfunction between female survivors and women in the general population.This is in contrast to studies from the United States (US), indicating that female survivors are at risk for impaired sexual functioning across domains [13,14] but corresponds with results of a Swedish study [11].Future studies are needed to determine if the differing results between our study and the US studies can be explained by differences in sampling, measures, definitions and/or cultural differences between the countries.For example, the definition of sexual dysfunction in one of the US studies [13] was based on the reports of the controls, making the definition sensitive to selection bias of the controls.The definition used in this study was, instead, based on standard scores with thresholds to identify clinically meaningful sexual dysfunction [25].Moreover, the estimates of dysfunction in our comparison group corresponds with previous results for general population samples [29,36,37], highlighting that problems with sexual function are common, in particular among women.
In contrast to our findings for females, male survivors were at increased risk of sexual dysfunction compared with men in the general population, with a higher prevalence of dysfunction related to Orgasm e ability and Erectile function.Thus, in line with a previous study [11], our results indicate that childhood cancer can have adverse effects on certain domains of sexual functioning, and that this is more evident in male survivors.It is well known that men and women differ in perceived sexual function and that sexual dysfunction is more prevalent in women [23,36].Therefore, analyses of survivors' sex life should be performed by gender and include appropriate noncancer comparison data, as in the present study.
Previous studies have shown conflicting results regarding risk of erectile dysfunction in childhood cancer survivors [9,10].Erectile dysfunction can be caused by a variety of reasons relating to psychological, endocrine, vascular and neurological factors [38].The increased risk of erectile dysfunction that we observed might be related to both physical and psychological impact of the cancer.Body image disturbance was associated with sexual dysfunction in several domains for male as well as female survivors, highlighting this issue as a target for interventions aiming at Statistically significant (P < 0.05) differences in the logistic regression multivariable models indicated in bold.a Self-reported dysfunction defined as cut-off Z 1 SD above/below the t-score mean of the norm population [24].
b Based on reports for the domains: Satisfaction with sex life, Interest in sexual activity, Orgasm e ability, Orgasm e pleasure, Vaginal lubrication (females only) and Erectile function (males only).improving sexual health.Results of an interview study with adult survivors indicate that a negative body image as a result of the cancer experience hinders physical intimacy [39], which can explain our findings.Furthermore, consistent with previous reports [9,13,15], we found emotional distress to be linked to sexual dysfunction.The associations between sexual dysfunction and body image disturbance and emotional distress, respectively, may be bidirectional and underscore that sexual health interventions should consider combining psychological (e.g.counselling) and medical (e.g.hormones) treatment.
Conflicting results have been presented as to how specific cancer treatments are associated with sexual dysfunction [12e15].Our findings demonstrate that the very or most intensive cancer treatments put survivors, particularly females, at increased risk for later sexual dysfunction.Patients treated with the very or most intensive treatments according to ITR-3.0 include those treated with hematopoietic stem cell transplantation (HSCT) and with relapse protocols, those treated for osteosarcoma, Ewing sarcoma and higher stages of other solid tumours, as well as those treated for acute myeloid leukaemia, acute lymphoblastic leukaemia with high-risk/T-cell lymphoma protocols and brain tumours with two or more modalities [35].Risk of gonadotoxicity is not specifically measured in the ITR-3.0rating system.However, cancer types in very or most intensive treatments groups are treated with high cumulative doses of alkylating chemotherapy, cranial or abdominal irradiation or HSCT, which are associated with increased risk of premature menopause, testicular dysfunction and deficiency of sexual hormones.Intensive treatment as a risk factor is an important observation that warrants further investigation to elucidate factors underpinning the higher risk observed among these survivors.
The strengths of the present study include the large population-based cohort, use of treatment information from a national quality registry (NQRCC) with good coverage, use of standardised measures and inclusion of a randomly selected comparison group.Nevertheless, some limitations should be recognised.Our estimates of dysfunction rely on self-reports and were not validated against, for example, structured interviews by professionals experienced in diagnosing sexual disorders.Furthermore, the definition of sexual dysfunction was based on American norms; however, aspects of sexual function and activity are similar between the countries [36,40], including age for sexual debut.Moreover, women and men in our comparison group report similar rates of dysfunction as the US general population [36].While our response rates are similar to Ref. [13] or higher [14] than other survey studies on this topic, there is still a risk of participation bias.Non-sexually active individuals might be less likely to participate in studies on sexuality, leading to sexual problems being underreported.Among male survivors, non-responders were older than responders, but we do not believe that the 1year difference resulted in an overestimation of dysfunction in this group.Finally, as the comparison group differed from survivors in some sociodemographic characteristics the analyses between these groups were adjusted accordingly.

Conclusions
Overall, young adult survivors of childhood cancer report sexual functioning in line with peers, but male survivors have more problems related to erectile function and orgasm.Furthermore, survivors are partnered and have children to a lesser extent, indicating difficulties in building intimate relationships.This underscores the need for routine assessment of sexual health in follow-up care of survivors.Survivors subjected to a more intensive cancer treatment and who experience concurrent psychological concerns are highrisk subgroups who may benefit from targeted screening and interventions.

Table 1
Sociodemographic, psychological and clinical characteristics of the study sample.

Table 2
Prevalence of sexual dysfunction a by sex in survivors and comparison group; differences between groups were adjusted for sociodemographics (age at study, education, relationship status, having children and country of birth).

Table 3
Results of the multivariable logistic regression models for sexual dysfunction in female survivors.The model for each SexFs domain included only the factors that were associated with sexual dysfunction in the specific domain in bivariate analysis.
BIS, Body Image Scale; CI, confidence interval; CNS, central nervous system; HADS, Hospital Anxiety and Depression Scale; OR, odds ratio; Ref, reference category.Statistically significant (P < 0.05) factors in the multivariable model indicated in bold.