Summary
Background and objectives The few existing studies of sexual dysfunction in women on hemodialysis are limited by small sample size. This large, cross-sectional study evaluated the prevalence and correlates of female sexual dysfunction in advanced kidney disease.
Design, setting, participants, & methods A total of 1472 women with ESRD undergoing hemodialysis were recruited to a multinational, cross-sectional study conducted within a collaborative dialysis network in Europe and South America. Sexual dysfunction was identified by the Female Sexual Function Index. Correlates of self-reported sexual dysfunction were identified by regression analyses.
Results Of the 1472 women, 659 completed questionnaires (45%). More than half (362 of 659 [55%]) lived with a partner, and 232 of 659 (35%) reported being sexually active. Of these 659 respondents, 555 (84%) reported sexual dysfunction. Women with a partner (282 of 362 [78%]) were less likely to report sexual dysfunction than those without a partner (273 of 297 [92%]) (P<0.001). Sexual dysfunction was independently associated with age, depressive symptoms, less education, menopause, diabetes, and diuretic therapy. Nearly all women who were not wait-listed for a kidney transplant and were living without a partner (249 of 260 [96%]) reported sexual dysfunction. More than half (128 of 232 [55%]) of sexually active women reported sexual dysfunction, associated with age, depressive symptoms, menopause, low serum albumin, and diuretic therapy.
Conclusions This descriptive study suggests most women on hemodialysis experience sexual problems. Additional research on the relevance of sexual dysfunction to symptom burden and quality of life in these women is needed.
Introduction
Hemodialysis, although life-preserving, is associated with poor survival, a high symptom burden, and impaired quality of life (1–4). Depression, pain, itch, impaired sleep, and fatigue are commonly reported by people undergoing long-term hemodialysis (5–7). Sexual dysfunction may also contribute to the symptom burden of CKD. Overall, three quarters of men on hemodialysis experience erectile dysfunction (8). Although phosphodiesterase-5 inhibitors improve erectile function in men with CKD, only a few small trials have evaluated such interventions (9).
Compared with the increasing awareness of erectile dysfunction in men on hemodialysis (8), sexual dysfunction in women with CKD is less well understood and researched. Studies reporting female sexual dysfunction in CKD are small (each with <150 participants) and report a wide variation in the prevalence of sexual difficulties (30%–100%) in these women (10–16). No randomized trial has evaluated interventions for female sexual dysfunction in CKD so far. A large descriptive study is now essential to understanding the prevalence, severity, and key correlates of sexual dysfunction in women with CKD to determine whether future research to evaluate health outcomes and screening and intervention strategies is warranted.
Materials and Methods
Population and Data Collection
We recruited women from 27 randomly selected hemodialysis clinics located in Europe and South America. The clinics are part of a collaborative dialysis network coordinated by Diaverum, which provides in-center hemodialysis for approximately 17,000 people in 14 countries. Clinics were selected using a computer-generated random-number sequence. Consecutive women age 18 years or older who were undergoing long-term hemodialysis between January and June 2008 were eligible. We approached all women in the participating clinics during hemodialysis treatment. Ethics approval was obtained from local ethics committees in each country in which the study was conducted. Women were enrolled after they provided written informed consent. The study was conducted according to the principles of the Declaration of Helsinki.
The presence of sexual dysfunction was assessed by the 19-item Female Sexual Function Index (FSFI) (17). This instrument evaluates six domains of sexual function in women in the 4 weeks before completion of the questionnaire: (1) desire, (2) arousal, (3) lubrication, (4) orgasm, (5) global satisfaction, and (6) pain (Supplemental Table 1). Each domain is given a maximum score of 6, providing a possible score range of 1.2–36 (Supplemental Table 2). A summary score less than 26.55 in the general population indicates female sexual dysfunction (18). We evaluated depressive symptoms concurrently using the Center for Epidemiologic Studies-Depression (CES-D) instrument (19). A score ≥18 in people on dialysis is compatible with depression (20). We used questionnaires in the participant’s native language after translation and linguistic validation by the MAPI Institute (http://www.mapi-institute.com). Questionnaires were completed by women anonymously and were collated with de-identified sociodemographic and clinical data provided by the treating physician on a standardized case report form.
Statistical Analyses
Baseline variables were calculated as mean ± SD or median and interquartile range (IQR) for continuous variables and as frequencies and percentages for categorical variables. We compared the characteristics of the women who provided responses to the FSFI questionnaire with those of nonrespondents using the chi-squared test and standard univariate statistics. Women were classified as respondents when they answered at least 18 questions on the FSFI questionnaire. Clinical characteristics of women with versus those without sexual dysfunction were compared using Pearson chi-squared tests and Mann-Whitney U tests.
We used stepwise multivariable logistic regression analysis to identify correlates of sexual dysfunction. We included the following in the multivariate model: age, depressive symptoms (CES-D score), presence of partner, occupational and menopausal status, smoking, housing, educational attainment, geographic region (Europe or South America), comorbid conditions (cardiovascular or cerebrovascular event, diabetes, or hypertension), cause of CKD (hypertension, diabetes, or other), neurologic conditions (spinal cord lesions, multiple sclerosis, Parkinson disease, or Alzheimer disease), previous kidney transplant, wait-listing for kidney transplant, medication (separately, β-blockers, angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, diuretics, erythropoietin, nitrates, lipid-lowering agents, antidepressants, antipsychotics, and anxiolytics), dry weight, interdialytic weight gain, time on dialysis, dialysis adequacy (expressed as Kt/V), dialysis blood flow, parity, and laboratory variables (hemoglobin, parathyroid hormone, calcium, phosphorus, albumin, and ferritin). Because of the likely relationship between age and occupational status and other potentially explanatory variables for sexual dysfunction, particularly the likelihood of having a partner and menopausal status, we conducted multivariate analyses excluding age and occupational status from the final model. We calculated the C-index for each multivariate model to measure how well the model discriminated between women who reported sexual dysfunction and those who did not. The C-index depicts a trade-off between the sensitivity and specificity of the model and ranges from 0.5 (no discrimination) to 1.0 (perfect discrimination).
Results are expressed as adjusted odds ratios (AORs) and their 95% confidence intervals (CIs). For all multivariable logistic analyses, continuous variables were categorized according to tertiles or specific clinically relevant threshold values. We then used the recursive partitioning and amalgamation method to identify combinations of clinical characteristics that were correlated with the presence of sexual dysfunction (21). Age served as the global correlate in our modeling. Finally, we performed a sensitivity analysis in the women who reported being sexually active. All statistical analyses were carried out using SAS software, version 9.1 (SAS Institute, Inc., Cary, NC). The recursive partitioning and amalgamation method analysis was conducted using an SAS macro routine (22).
Results
Participants
A total of 1591 women were eligible for the study. We excluded 42 women because they were enrolled in other studies, and another 77 declined to participate. The remaining 1472 (93%) were enrolled; 659 women (45%) completed and returned questionnaires. Individuals who did not respond (813 women; 55%) were older, were less likely to be living with a partner or to be wait-listed for a kidney transplant, and were more likely to have reached menopause (P≤0.001 for all).
Prevalence of Sexual Dysfunction
Overall, 555 women (84.2%) on hemodialysis reported FSFI scores consistent with sexual dysfunction (score < 26.55). The median score was 14.8 (IQR, 2.0–33.1). Characteristics of women according to the presence of sexual dysfunction are reported in Table 1. Individual domain scores were (in order from most to least dysfunction [maximum score of 6]): arousal, median score of 1.8 (IQR, 0–5.4); orgasm, 2 (IQR, 0–6); desire, 2.4 (IQR, 1.2–4.8); lubrication, 2.4 (IQR, 0–6); pain, 3.4 (IQR, 0–6); and satisfaction, 3.6 (IQR, 0.8–6). Overall, 282 (77.9%) women with a partner and 273 (91.9%) women without a partner reported sexual dysfunction (P<0.001).
Demographic and clinical characteristics of women undergoing hemodialysis according to presence of sexual dysfunction
Correlates of Sexual Dysfunction
On univariate analyses, women who reported sexual dysfunction were older, had lower education attainment, and had more depressive symptoms (Table 1). European women were more likely than women in South America to report symptoms of sexual dysfunction. Women with sexual dysfunction were less likely to be living with a partner, to be wait-listed for a kidney transplant, and to have had a previous kidney transplant or be employed and were more likely to have children, to be menopausal, and to have diabetes. Women reporting sexual dysfunction had also been on dialysis for less time, had lower dialysis adequacy, had a higher body weight, and were more likely to be prescribed diuretic or nitrate therapy. Hemoglobin levels, cardiovascular disease, and hypertension did not correlate with sexual dysfunction.
In multivariate analysis, age correlated strongly with sexual dysfunction; prevalence increased 8% (AOR, 1.08 [95% CI, 1.06–1.11]) per 1-year increase. Women who were not wait-listed for a transplant, received a pension (AOR, 3.02 [95% CI, 1.40–6.51]) or were unemployed (AOR, 2.56 [95% CI, 1.18–5.57]), had a CES-D score consistent with depression (AOR, 2.71 [95% CI, 1.55–4.73]), had a higher interdialytic weight gain (AOR, 2.11 [95% CI, 1.04–4.27]), or were unmarried (AOR, 4.55 [95% CI, 2.38–8.33]) had an increased risk for sexual dysfunction. Because of a likely relationship between age and occupational status and other potential explanatory variables, we also conducted multivariate analysis excluding these variables (Figure 1). In this analysis, having a partner or being wait-listed for a kidney transplant correlated with a lower risk for sexual dysfunction, whereas depressive symptoms, lower educational attainment, reaching menopause, having diabetes as a cause of kidney disease, and receiving a diuretic independently correlated with an approximately two- to four-fold increased risk for sexual dysfunction. The C-index was 0.89, suggesting that this model discriminated well between women who did and those who did not experience sexual dysfunction.
Sociodemographic and clinical correlates of self-reported female sexual dysfunction (n=659), displayed as multivariate-adjusted odds ratios. The multivariate model included depressive symptoms (Center for Epidemiologic Studies-Depression score [CES-D]), partner and menopausal status, smoking, housing, educational attainment, geographic region (Europe, South America), comorbid conditions (cardiovascular or cerebrovascular event, diabetes, or hypertension), cause of kidney disease (hypertension, diabetes, or other), neurologic conditions (spinal cord lesions, multiple sclerosis, Parkinson disease, or Alzheimer disease), previous kidney transplant, wait-listing for kidney transplant, medication, dry weight, interdialytic weight gain, time on dialysis, dialysis adequacy (expressed as Kt/V), dialysis blood flow, parity, and laboratory variables (including hemoglobin, parathyroid hormone, calcium, phosphorus, albumin, and ferritin).
Recursive partitioning and amalgamation method analysis (Figure 2) identified six groups of women with distinct risks for sexual dysfunction. This analysis identified being wait-listed for a kidney transplant as the most important discriminating variable correlating with sexual dysfunction. We then assigned the reference category as the group of women with the lowest risk for sexual dysfunction (premenopausal women on a kidney transplant waiting list; assigned AOR, 1.00). Women without a partner who were not wait-listed for a kidney transplant had the highest risk for sexual dysfunction (AOR, 21.91 [95% CI, 9.99–48.04]).
Identification of subgroups of women requiring hemodialysis who had different risks for sexual dysfunction: results of recursive partitioning and amalgamation method analysis. A tree-growing algorithm modeled age-adjusted odds ratios (AORs) for predicting risk for self-reported sexual dysfunction. The splitting variables are shown between branches, and the characteristic assigning women to different groups is above the corresponding branch. Premenopausal women wait-listed for a kidney transplant had the lowest prevalence of sexual dysfunction and were designated as the reference category. Circles indicate subgroups of patients. Shaded squares indicate women with different combinations of clinical and sociodemographic characteristics and the associated correlation with sexual dysfunction (e.g., women without a partner who are not wait-listed for a kidney transplant had an adjusted risk of 21.91 compared with the lowest-risk category of premenopausal women wait-listed for a transplant). Numbers inside circles and squares represent numbers of patients with (upper number; dark border) and without (lower number) sexual dysfunction, respectively.
Sensitivity Analyses
Because sexual dysfunction scores are influenced by sexual activity in the FSFI questionnaire, we assessed the prevalence and correlates of sexual dysfunction in women who were sexually active in the 4 weeks before completion of the questionnaire. Of the 232 respondents (35%) reporting sexual activity, 128 (55.2%) reported sexual dysfunction. The median FSFI score in sexually active women was 25.8 (IQR, 22.0–30.0). Univariate analysis revealed similar correlates of sexual dysfunction compared with those in the overall study population. Sexually active women with sexual dysfunction were older (mean age, 51.7±14.0 years versus 44.6±12.1 years among those without sexual dysfunction); had higher CES-D scores; were more likely to receive a pension, be menopausal, take nitrate therapy, and have diabetic nephropathy or lower dialysis adequacy; and were less likely to be wait-listed for a transplant (P<0.05). On multivariate analyses, age correlated with sexual dysfunction (1-year increase associated with 5% increase in prevalence) (AOR, 1.05 [95% CI, 1.03–1.07]), CES-D score, and low serum albumin. When age was excluded, sexual dysfunction in sexually active women correlated with CES-D score, menopause, low serum albumin, and diuretic therapy (Figure 3). The C-index for the model’s ability to discriminate between sexually active women with and those without sexual dysfunction was 0.77.
Sociodemographic and clinical correlates of self-reported female sexual dysfunction in sexually active women (n=232), displayed as multivariate-adjusted odds ratios. The multivariate model included depressive symptoms (Center for Epidemiologic Studies-Depression score [CES-D]), partner and menopausal status, smoking, housing, educational attainment, geographic region (Europe, South America), comorbid conditions (cardiovascular or cerebrovascular event, diabetes, or hypertension), cause of kidney disease (hypertension, diabetes, or other), neurologic conditions (spinal cord lesions, multiple sclerosis, Parkinson disease, or Alzheimer disease), previous kidney transplant, wait listing for kidney transplant, medication, dry weight, interdialytic weight gain, time on dialysis, dialysis adequacy (expressed as Kt/V), dialysis blood flow, parity, and laboratory variables (including hemoglobin, parathyroid hormone, calcium, phosphorus, albumin, and ferritin).
Discussion
According to a questionnaire-based definition, sexual dysfunction was highly prevalent in women undergoing hemodialysis. Four of five women having in-center hemodialysis reported sexual dysfunction. Sexual problems were associated with age, depressive symptoms, menopause, low educational attainment, and diabetes and were nearly always present in women who were not wait-listed for a kidney transplant. Having a partner was associated with less sexual dysfunction. Our results indicated that although more than half of women had a partner, only one third of women on hemodialysis were sexually active during the study period. Approximately half of sexually active women on hemodialysis reported sexual problems, particularly those who had reached menopause or had depressive symptoms, had low serum albumin, or had received diuretic therapy.
Previous data on the prevalence and correlates of sexual dysfunction in women with CKD are sparse. Five studies (n=382 women), all published since 2005, have evaluated sexual function in women requiring peritoneal dialysis or hemodialysis and report a prevalence of 30%–100% (10,12,13,15,16). Together, these studies had limited capacity to identify the independent correlates of sexual dysfunction in women on dialysis because of small sample sizes. The study reporting the lowest prevalence of female sexual dysfunction (30%) (16) used the Index of Female Sexual Function (23) to assess sexual function instead of the FSFI used in the other available studies; thus, the overall prevalence of sexual difficulty in that study may have been underestimated. Our larger study suggests that the overall prevalence of sexual dysfunction in women on dialysis may be considerably higher, irrespective of sexual activity, and that depressive symptoms are strongly associated with sexual dysfunction.
Our finding that 84% of women on hemodialysis and 55% of sexually active women on hemodialysis experience sexual dysfunction is higher than rates reported in the general population in the United States, of whom approximately 40% who are sexually active experience sexual difficulty (24,25). Although data on prevalence of sexual dysfunction in women with other chronic diseases are limited by small sample sizes and are heterogeneous with respect to the assessment instrument used and inclusion of women who are not sexually active, the prevalence of sexual dysfunction in women on dialysis may also exceed that in women with other chronic conditions, including metabolic syndrome (38% affected) (26), type 1 diabetes mellitus (27), rheumatoid arthritis (45%) (28), and multiple sclerosis (35%) (29). Sexual dysfunction rates in women on hemodialysis may, however, be similar to those in sexually active women of a similar age in the general population (40–59 years); of these, approximately 55% fulfill criteria for sexual dysfunction (30) and 60%, for coronary artery disease (31). The prevalence of sexual dysfunction in women who are of an age similar to that of our cohort and who have normal kidney function remains incompletely researched. As was found in our study, existing observational studies in the wider community population also identify age, marital status, parity, postmenopausal status, educational attainment, and poor overall health as predictors of sexual function (24,30,32). Similar to our finding that arousal and orgasm are the most severely affected aspects of sexual health in women on dialysis, the most commonly observed sexual problems in women in the general population are lack of interest in sex, inability to orgasm, and reduced desire (24).
The relevance of finding sexual dysfunction in most women on hemodialysis is unclear. The cutoff value of 26.55 in the FSFI to identify sexual dysfunction was developed in the general population (18); whether using this index in women with CKD of any severity is clinically valid requires additional qualitative research. Validation of the FSFI in non–general population datasets has not been published to our knowledge. In addition, consequences of finding sexual dysfunction in our study remain speculative, particularly because we did not ask women about the effects of sexual dysfunction on their well-being. Indeed, although the FSFI is considered a valid tool for identifying sexual dysfunction in women (18), the index itself does not include questions about whether sexual problems are associated with psychological distress (17). Understanding the subjective experience of sexual functioning in these women is essential before approaches to treatment can be considered. In broad community populations, 35%–40% of women with sexual problems report being dissatisfied or very dissatisfied with their overall sex lives (25).
Outcome data from randomized trials in the general population on treatment of sexual dysfunction in women are unconvincing because they rely on surrogate outcomes, such as genital engorgement, changes in sexual function scores, and changes in female sexual function after treatment of their male partner (33–36). The clinical relevance of such outcomes to women is not clear. Our systematic review in CKD found no randomized trials that evaluated interventions for sexual dysfunction in women (9 ); however, given the uncertainty regarding the health consequences of sexual dysfunction in this population, the absence of such trials may be entirely appropriate. A randomized trial of two management strategies for the implementation of treatment for pain, sexual dysfunction, and depression in patients undergoing long-term hemodialysis (the Symptom Management Involving End-Stage Renal Disease [SMILE] study) has completed recruitment (37). In this trial, women who report the presence of sexual dysfunction and an interest in learning about treatment options have been recommended to have a gynecologic evaluation, although the investigators acknowledge that an evidence-based approach to treat sexual dysfunction in women on hemodialysis is not established (37).
Although we report the largest study of sexual function in women on hemodialysis to date, this study has limitations that should be considered. We did not include data for race or ethnicity, which has been shown to correlate with sexual function in other populations (24). We had a response rate of only 45%. This response is similar to that of other studies of sexual health in CKD (8), but women who did not respond were older, were less likely to be married, and were more likely to be menopausal. Older women were probably less willing to discuss sexual issues with healthcare providers, even under the assurance of anonymity, although they may be equally or more likely to experience sexual problems compared with their younger peers (38). The low response rate and important differences between respondents and nonrespondents may reduce the generalizability of our findings to all women on hemodialysis. Because of the cross-sectional design, we do not know whether sexual problems were persistent in our population. Sexual inactivity results in lower scores for many items in the FSFI; thus, including sexually inactive women in the analyses may have overestimated the prevalence of sexual dysfunction in our population. However, we included these women in our overall analyses because, as a result of the cross-sectional design, we could not determine whether sexual inactivity was a cause or a consequence of sexual problems.
Finally, observational studies of this nature are prone to confounding by both measured and unmeasured variables. As such, although not being wait-listed for a kidney transplant was an important independent correlate of sexual function after adjustment for other clinical variables in our population, it is unlikely that transplantation status was a direct cause or effect of sexual problems. Rather, eligibility for transplantation probably reflected overall well-being affected by numerous health variables that were the true link between transplant status and sexual function. Nevertheless, dialysis providers and future researchers might benefit from knowing that sexual dysfunction problems may be nearly universal in women on dialysis not wait-listed for transplantation. Similarly, this study showed that sexual dysfunction was associated with diuretic therapy, a treatment more commonly used in women who may have poorer fluid removal because of heart disease or hypotension (39,40). It is entirely probable that diuretic treatment does not contribute to sexual dysfunction directly but reflects the presence of comorbidity that could not be accounted for by multivariable adjustment.
In conclusion, sexual dysfunction may be common in women requiring hemodialysis and is linked to older age, ineligibility for kidney transplantation, not having a partner, depressive symptoms, lower educational attainment, and menopause. Our descriptive study is the first large study to examine the prevalence and correlates of female sexual function on hemodialysis and provides data to support the need for future analytical studies in this area (41). Given the potentially high prevalence of sexual problems in women on dialysis, quantitative and qualitative studies are required to understand the relevance of sexual dysfunction on overall quality of life and psychological well-being in these women. Validating indices of female sexual dysfunction in women with CKD is now necessary to evaluate whether existing questionnaires measure clinically relevant sexual problems in this population. Such additional studies will also help determine whether funding research on new screening and intervention strategies for sexual dysfunction in women on hemodialysis is warranted.
Disclosure
None.
Acknowledgments
Amgen Inc. provided an unconditional independent research grant of €10,000 for administrative expenses of this study. The study was independently run by the Mario Negri Sud Consortium. Suetonia Palmer received funding from an Amgen Dompé Consorzio Mario Negri Sud Fellowship.
The funding bodies had no role in study design, collection, analysis and interpretation of data, writing of the report, or decision to submit the report for publication.
This research was presented at the American Society of Nephrology Kidney Week 2011, November 10–13, 2011, Philadelphia, Pennsylvania (FR-PO1681).
The members of the Writing, Steering, Executive Committee and the Medical Committee of the Collaborative Depression and Sexual dysfunction (CDS) in Hemodialysis Working Group are listed in the Supplemental Materials.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.12601211/-/DCSupplemental.
See related editorial, “Female Sexual Dysfunction in ESRD: An Underappreciated Epidemic?,” on pages 881–883.
Access to UpToDate on-line is available for additional clinical information at www.cjasn.org.
- Received December 12, 2011.
- Accepted March 12, 2012.
- Copyright © 2012 by the American Society of Nephrology