Childhood Sexual Abuse and Pelvic Floor Dysfunction in Midlife Women in the Study of Women's Health Across the Nation

Abstract

Objectives:

Research has shown a link between childhood sexual abuse (CSA) and lower urinary tract and sexual disorders in clinical settings. We examined whether CSA was associated with two specific aspects of high tone, elevated resting tension pelvic floor dysfunction (PFD) in community-dwelling women.

Materials and Methods:

Data were from 2068 participants (25.5% Black, 9.6% Chinese, 10.8% Japanese, 5.0% Hispanic, and 49.1% Non-Hispanic White) in the Study of Women's Health Across the Nation (SWAN), a multirace/multiethnic longitudinal observational study of women's midlife health. At baseline, enrolled women were 42–52 years old and premenopausal or early perimenopausal. Annual or biennial assessments conducted over 20 years (1996 through 2017) included single-item queries about urgency urinary incontinence and pain with sexual activity used to assess PFD outcomes. The 12th follow-up visit conducted in 2009–2011 assessed the primary exposure, history of CSA, using a single-item response. Multivariate logistic regression models tested study objectives.

Results:

The prevalence of CSA was 15%, self-reported in 313/2068 women. CSA and PFD, both pain with sexual activity (odds ratio [OR] = 1.56 confidence interval [95% CI = 1.12–2.18]) and urgency urinary incontinence (OR = 1.87 [95% CI = 1.29–2.71]), were significantly associated in unadjusted models. The final adjusted model that included sociodemographic variables and physical and behavioral risk factors was significant for pain with sexual activity (OR = 1.48 [95% CI = 1.08–2.02]), but not for urgency urinary incontinence (OR = 1.38 [95% CI = 0.96–1.98]).

Conclusions:

In midlife women, pain with sex, but not urgency urinary incontinence, was associated with a history of CSA. A multidisciplinary diagnostic and therapeutic approach to PFD is key, inclusive of CSA screening.

Introduction

Sexual abuse history is common in women, with reported rates ranging from 3% to 25%.^1,2 Epidemiological studies show associations between early childhood sexual abuse (CSA) and subsequent psychological and physical health issues, including obesity.^3–5 In particular, childhood trauma-related Post Traumatic Stress Disorder and depression have been associated with physical comorbidities.^6,7 In midlife women, childhood sexual trauma has been associated with increased risk of obesity.^7,8 In the Study of Women's Health Across the Nation (SWAN), analyses showed a significantly higher waist circumference and body mass index (BMI) at baseline in women with a history of any childhood abuse/neglect, and specifically physical and sexual abuse, compared to women with no childhood abuse/neglect history.⁹ Prior childhood abuse/neglect, particularly childhood sexual trauma, affects adult women's health.

A community-based study found women in the general population with up to three common gynecological complaints (dysmenorrhea, menorrhagia, and/or sexual dysfunction) were at higher risk of reporting a history of sexual assault, the risk being highest in women with all three complaints.¹⁰ A clinical study found that multiple pelvic floor dysfunctions (PFDs), including sexual, urinary, and fecal complaints, were reported in women presenting to a gynecology clinic for pelvic pain assessment.^11,12 Overactive bladder symptoms have also been associated with exposure to sexual abuse in women presenting for medical care in Germany, but not in a study of women presenting for urogynecological care in New Mexico.^13,14 Recently, the link between sexual trauma exposure and PFD, in particular lower urinary tract, was found to vary by race/ethnicity in a biracial cohort of Latina and White women.¹⁵

We evaluated high-tone PFDs using available questionnaire responses from the SWAN. High-tone dysfunction includes elevated resting muscle tension and pain, with reduced muscle relaxation and subsequent dysfunctions.^16,17 SWAN provides us with the opportunity to explore the associations between childhood sexual trauma and two specific aspects of high-tone PFD, urgency urinary incontinence and pain with intercourse, in participants in a multiracial/ethnic cohort of community-dwelling midlife women.

Materials and Methods

Participants

Participants were from the SWAN, a multisite, multiethnic/multiracial, longitudinal, prospective study of women's health and aging through the menopausal transition, conducted at seven sites in the United States.¹⁸ Inclusion criteria were ages 42–52 years, not pregnant or breastfeeding, an intact uterus and at least one ovary, reported menstrual bleeding within the prior 3 months, not currently using medications known to affect pituitary or ovarian function, and had not used exogenous hormones within the 3 months preceding the baseline interview. Following the baseline interview, participants completed up to 15 follow-up interviews that occurred annually or biannually. Assessments included self-administered and interviewer-administered questionnaires assessing social, economic, behavioral, psychological, health, and lifestyle characteristics.

Several population-sampling techniques were used and IRB approval was obtained by the seven sites, as previously described, and women provided informed consent to participate. The baseline examination, conducted between 1996 and 1997, enrolled 3302 women from 5 racial/ethnic groups, including Non-Hispanic White, Black, Japanese, Chinese, and Hispanic. Each site enrolled Non-Hispanic White participants and participants of one other race/ethnicity. Interviews and questionnaires were available in English, Spanish, Cantonese, and Japanese.

Study variables

Data on CSA were obtained using a single item from the sexual abuse subscale of the Childhood Trauma Questionnaire (CTQ) that was included in the 12th follow-up (F/U 12) visit (2009–2011): “As a child were you ever sexually attacked, raped, or sexually abused?”¹⁹ While there are five subscales in the CTQ-emotional abuse, physical abuse, sexual abuse, neglect, and minimization/denial, our analysis focused on the dimension of sexual abuse only. We used the single question available at the F/U 12 response to maximize sample size. The short-form CTQ was administered at F/U 15 enabling us to determine the association between the CTQ scores and the single-item response by conducting a Wilcoxon nonparametric test.

In our analyses, there was a statistically significant difference between women who responded “yes” to the single question versus those who responded “no” with a higher median score on the CTQ subscale for the former group (median = 12) versus lower median score (median = 5) for the latter group (p < 0.0001), revealing a strong association between the two constructs. This aligns with the previously established correlation between the single question and the CTQ sexual abuse subscale (r = 0.62, p < 0.0001).²⁰

Data for the pelvic floor outcomes included the following: (1) urgency urinary incontinence asked at baseline and each FU visit, “In the last month have you lost urine, even a small amount, beyond your control when you have the urge to urinate and can't get to the toilet fast enough?” and (2) Pain with intercourse asked at most visits (all, except FU visits 7, 9, 11, and 14), “During the past 6 months have you felt pelvic or vaginal pain during intercourse?” dichotomized to use as Yes/No variable, grouping participants who responded “always,” “almost always,” and “sometimes” as Yes versus “almost never” and “never” as No. Participants who reported no intercourse for the past 6 months were excluded in the models with the pain with intercourse outcome. Note the International Continence Society has recently changed terminology from urge urinary incontinence, original terminology in SWAN, to urgency urinary incontinence.²¹

Covariates collected at baseline included site; race/ethnicity (self identification as Black, Chinese, Hispanic, Japanese, or Non-Hispanic White); marital status; age; number of vaginal births, and education (achieving less than a high school diploma, high school diploma, some college, college degree, through post-graduate education).

Time-varying covariates collected for each FU visit included the following: vaginal dryness, menopausal status (premenopausal [no bleeding irregularity in past 3 months], early perimenopausal [less predictable menses in last 3 months], late perimenopausal [no menstrual bleeding for at least 3 months but no >12 months], post menopausal [no menstrual bleeding for at least 12 months], and undetermined [use of hormone therapy or hysterectomy without bilateral oophorectomy before 12 months of amenorrhea]); BMI (weight per stadiometer in kilograms divided by height by calibrated scale in meters squared); depressive symptoms (≥16 on Center for Epidemiological Studies Depression); and summed anxiety score.^22–24 Covariates were based on the literature and biological plausibility for confounding the main association of interest.

Statistical methods

All variables were examined using descriptive statistics: means and standard deviations for continuous variables and frequencies and percentages for categorical variables. All variables were checked for normality or possible outliers. If needed, variables were transformed accordingly. Bivariate associations of all independent variables were evaluated. Site was included as a random effect variable to adjust for participants clustering within site.

Two different sets of analyses explored each outcome separately. Samples varied slightly with N = 2038 in analyses of urge urinary incontinence (UUI) and N = 1812 in analyses of pain since respondents who were not sexually active over the past 6 months were excluded from the analyses for painful intercourse. The statistical modeling approach was the same for each outcome.

We conducted a series of longitudinal mixed effects logistic regressions with urgency urinary incontinence as a longitudinal binary outcome from visits 0 through 15 and childhood sexual trauma as the primary predictor. To account for the correlation between repeated measurements for each participant, we used the GLIMMIX procedure, which incorporates the appropriate covariance structure, random effects, and a binary outcome variable. All models include the effect of time, which measures changes in the response variable over time when all other variables are kept constant. In all models, clinical site was included as a random effect to account for clustering based upon site.

The base model (model 0) included childhood sexual trauma, time, and race/ethnicity. Model 1 is an extension of model 0+site+baseline sociodemographic variables (age, marital status, and education). Model 2 is an extension of model 1+number of vaginal births at baseline and time-varying physical and behavioral risk factors (BMI, CESD score, anxiety score, and menopausal status). We repeated the above analysis for the outcome pain during intercourse. We included race/ethnicity-by- childhood sexual trauma and time-by-childhood sexual trauma interactions to examine whether the association between trauma and PFD differed by race/ethnicity or over time, respectively. In supplementary analysis, we re-analyzed the final model for the urgency urinary incontinence outcome eliminating BMI as a covariate. All analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC).

Results

The participants were 45.9 (±2.7) years of age at baseline, and 49% identified as Non-Hispanic White. The prevalence of CSA was 15%, self-reported in 313/2068 women. About two thirds were married or living as married (Table 1, total N = 2038). Women who experienced CSA were more likely to be Non-Hispanic White or Black, separated/widowed/divorced, and from the Michigan site. They had higher BMI, CESD, and anxiety scores than women who did not report this history (Table 1, p < 0.001 for all comparisons). In total, 30% of participants (626/2068) never reported UUI at any visit, 7% (150/2068) reported UUI at all visits, and 63% (1292/2068) had alternating responses. Pain with intercourse was never reported by 41% of participants (738/1812), always reported by 6% of participants (108/1812), and reported at some visits by 53% of participants (966/1812).

Table 1.

Participant Characteristics at Baseline by Childhood Sexual ABUSE status

Participant characteristics	Total sample, N = 2068	No CSA, N = 1755 (84.9%)	CSA, N = 313 (15.1%)	p
Age at baseline (years), mean (SD)	45.9 (2.7)	45.9 (2.7)	45.6 (2.6)	0.08
Race/ethnicity, N (%)
Black	528 (25.5)	425 (24.2)	103 (32.9)	<0.0001
Non-Hispanic White	1,015 (49.1)	846 (48.2)	169 (54.0)
Chinese	198 (9.6)	183 (10.4)	15 (4.8)
Hispanic	104 (5.0)	91 (5.2)	13 (4.2)
Japanese	223 (10.8)	210 (12.0)	13 (4.2)
Marital status, N (%)
Single/never married	278 (13.6)	231 (13.4)	47 (15.1)	0.002
Currently married/living as married	1396 (68.5)	1208 (70.0)	188 (60.5)
Divorced/separated/widowed	364 (17.9)	288 (16.7)	76 (24.4)
Education, N (%)
≤High school	426 (20.8)	367 (21.1)	59 (18.9)	0.09
>High school/some college	665 (32.4)	547 (31.4)	118 (37.8)
≥College	959 (46.8)	824 (47.4)	135 (43.3)
Site, N (%)
Michigan	348 (16.8)	274 (15.6)	74 (23.6)	0.0004
Boston	297 (14.4)	241 (13.7)	56 (17.9)
Chicago	245 (11.9)	207 (11.8)	38 (12.1)
UC Davis	352 (17.0)	304 (17.3)	48 (15.3)
UCLA	391 (18.9)	354 (20.1)	37 (11.8)
New Jersey	166 (8.0)	143 (8.2)	23 (7.4)
Pittsburgh	269 (13.0)	232 (13.2)	37 (11.8)
Body mass index (kg/m2), mean (SD)	27.7 (7.2)	27.4 (7.0)	29.2 (8.0)	0.0002
Number vaginal deliveries, mean (SD)	1.7 (1.6)	1.7 (1.4)	1.7 (1.5)	0.97
Depression (CESD ≥16), N (%)	468 (22.7)	375 (21.4)	93 (29.7)	0.001
Anxiety, mean (SD)	6.4 (2.3)	6.3 (2.2)	6.8 (2.7)	0.001
Number vaginal deliveries, median (25th, 75th)	2.0 (0, 3)	2.0 (0, 3)	2.0 (0, 3)	0.97
Menopausal status, N (%)
Premenopausal	1142 (55.5)	992 (56.8)	150 (48.1)	0.01
Perimenopausal	913 (44.4)	751 (43.0)	162 (51.9)
Undetermined/HRT	3 (0.2)	3 (0.2)	0

CSA, childhood sexual abuse.

CSA was statistically significantly associated with urgency urinary incontinence (Table 2) in unadjusted analysis (odds ratio [OR] = 1.87 confidence interval [95% CI = 1.29–2.71]). The finding persisted after adding baseline sociodemographic variables age, marital status, and education (MODEL 1) (OR = 1.70 [95% CI = 1.17–2.47]), but not after adding number of vaginal births and time-varying physical and behavioral risk factors BMI, CESD score, anxiety score, and menopausal status (MODEL 2) (OR = 1.38 [95% CI = 0.96–1.98]).

Table 2.

Risk of URGENCY Urinary Incontinence by Childhood Sexual ABUSE

Parametera	Model 0		Model 1		Model 2
	OR	95% CI	OR	95% CI	OR	95% CI
CSA	1.87	1.29–2.71	1.70	1.17–2.47	1.38	0.96–1.98
Years	1.10	1.09–1.10	1.09	1.09–1.10	1.09	1.08–1.10
Race/ethnicity
Non-Hispanic White	Ref.		Ref.		Ref.
Black	1.52	1.10–2.09	1.50	1.07–2.11	1.08	0.77–1.52
Chinese	0.23	0.14–0.37	0.22	0.14–0.37	0.32	0.20–0.53
Hispanic	0.20	0.10–0.38	0.26	0.12–0.52	0.20	0.10–0.40
Japanese	0.24	0.15–0.38	0.23	0.14–0.37	0.34	0.21–0.54
Age (years)			1.01	0.96–1.06	0.10	0.95–1.05
Marital status
Never			0.94	0.58–1.52	1.21	0.74–1.95
Married/living as married			0.86	0.60–1.25	0.89	0.62–1.27
Divorced/separated/widowed			Ref.		Ref.
Education
≤High school			0.79	0.54–1.15	0.61	0.42–0.89
>High school/some college			0.90	0.66–1.24	0.83	0.61–1.13
≥College			Ref.		Ref.
Body mass index (kg/m2)					1.07	1.05–1.08
Vaginal dryness					1.07	1.02–1.13
Depression (CESD ≥16)					1.04	0.89–1.21
Anxiety					1.10	1.06–1.13
No. of vaginal deliveries					1.16	1.05–1.29
Menopausal status
Premenopausal					Ref.
Perimenopausal					1.33	1.11–1.59
Postmenopausal					1.23	0.98–1.54
Undetermined/HRT					1.38	1.08–1.77

All models.

^a Site was included as random variable, N = 2068.

CSA was statistically significantly associated with pain with intercourse (Table 3) in unadjusted analysis (OR = 1.56 [95% CI = 1.12–2.18]). This finding persisted after adding baseline sociodemographic variables age, marital status, and education (MODEL 1) (OR = 1.58 [95% CI = 1.12–2.20]), and after adding number of vaginal births and time-varying physical and behavioral risk factors BMI, CESD score, anxiety score, and menopausal status (MODEL 2) (OR = 1.48 [95% CI = 1.08–2.02]). In Model 2, statistically significant covariates included Chinese and Japanese race/ethnicity; married or living as married; ≤high school education; vaginal dryness; CESD ≥16; anxiety score; and menopausal status.

Table 3.

Risk of Pelvic Pain During Intercourse by Childhood Sexual ABUSE

Parametera	Model 0		Model 1		Model 2
	OR	95% CI	OR	95% CI	OR	95% CI
CSA	1.56	1.12–2.18	1.58	1.12–2.20	1.48	1.08–2.02
Years	1.11	1.10–1.12	1.11	1.10–1.13	1.03	1.01–1.05
Race/ethnicity
Non-Hispanic White	Ref.		Ref.		Ref.
Black	0.79	0.59–1.06	0.74	0.53–1.01	1.12	0.82–1.53
Chinese	3.28	2.20–4.87	2.78	1.86–4.16	2.65	1.82–3.84
Hispanic	1.42	0.81–2.48	1.08	0.59–1.96	1.08	0.61–1.89
Japanese	2.42	1.65–3.56	2.22	1.51–3.27	2.19	1.53–3.15
Age (years)			0.98	0.94–1.03	0.97	0.93–1.01
Marital status
Never			1.43	0.85–2.42	1.19	0.73–1.96
Married/living as married			1.61	0.85–2.42	1.58	1.12–2.22
Divorced/separated/widowed			Ref.		Ref.
Education
≤High school			1.61	1.16–2.23	1.82	1.33–2.49
>High school/some college			1.14	0.87–1.51	1.17	0.90–1.52
≥College			Ref.		Ref.
Body mass index (kg/m2)					0.96	0.93–1.01
Vaginal dryness					2.04	1.92–2.17
Depression (CESD ≥16)					1.51	1.25–1.83
Anxiety					1.06	1.03–1.10
No. of vaginal deliveries					0.76	0.70–0.83
Menopausal status
Premenopausal					Ref.
Perimenopausal					1.36	1.10–1.68
Postmenopausal					2.77	2.09–3.66
Undetermined/HRT					1.93	1.44–2.57

All models.

^a Site was included as random variable, N = 1812.

In supplementary analysis, re-analyzing the final model for the urgency urinary incontinence outcome eliminating BMI resulted in a statistically significant correlation; adding BMI back to the model, the statistical significance is marginal (p = 0.08) with the previously reported OR and CI in Model 2. In addition, neither the race/ethnicity by CSA nor time by CSA interactions were statistically significant for either urgency urinary incontinence or the pain with intercourse outcome.

Discussion

This study provides additional insight into the relationship between CSA and PFD in midlife community-dwelling women. In multivariable models, childhood sexual trauma was significantly associated with PFD, specifically pain with intercourse, but not with urgency urinary incontinence, in this multiracial/multiethnic cohort of midlife women.

PFDs are common in midlife women with linear increases in urinary incontinence severity observed with age in American and Norwegian cohorts.^25,26 In a clinic-based study, women who presented with pelvic pain were more likely to have multiple PFDs involving urological, vaginal, and colorectal functions.¹² Previous studies have linked sexual trauma history with both lower urinary tract issues and pain with intercourse in women presenting for clinical care.^10,27 Clinical presentation differed in women patients with interstitial cystitis/painful bladder, with patients with sexual abuse history reporting greater pain, but fewer voiding complaints than patients without sexual abuse history.²⁸ A recent case–control study found that adverse exposure events, including abuse, may lead to overactive bladder and interstitial cystitis/painful bladder.¹⁷

The Lower Urinary Tract Dysfunction research network (LURN) performed secondary analysis of their observational cohort study of midlife men and women, treating for urinary symptoms, but not pelvic pain, from participating Urology and Urogynecology clinics, and found an association of childhood sexual trauma and urinary incontinence symptom severity.²⁹ Using community-based longitudinal data from a diverse population of 30–79-year-old men and women in the Boston area community health survey, researchers found a significant association between childhood and adolescent/adult sexual abuse and urinary frequency, urgency, and nocturia, and present a strong case for causality.³⁰

Characterizing PFDs in midlife is important as they can significantly impact quality of life. While urinary incontinence (UI) is often thought of as a problem for women in late life, 65 years of age and older, researchers from the Seattle Midlife Women's Health Study identified correlates of urinary incontinence, both stress and urgency, in midlife women to better understand a lifespan view of UI.³¹ They concluded that having a high BMI was correlated with UUI in midlife. Due to the correlation of high BMI with both our predictor CSA and outcome variable UUI, we re-analyzed the final model eliminating BMI and found that CSA and UUI were statistically significantly correlated. After adding BMI back to the model, the statistical significance is marginal. BMI mediates the relationship between CSA and UUI.

Strengths of the study include inclusion of a multiethnic/multiracial sample of community-dwelling women with questionnaires offered to individuals in their preferred language and prospective documentation of these pelvic floor symptoms throughout midlife with follow-up of outcomes for up to 20 years, while accounting for sociodemographic, physical, and behavioral risk factors. However, limitations of the study include the retrospective report of childhood abuse. We used a single item as predictor variable that was highly correlated with the validated childhood trauma questionnaire; we had more available data for the analyses with the single item.²⁰

In addition, the pelvic floor outcome for urgency urinary incontinence used a single nonvalidated question. It queries about the last month of symptoms only and may oversimplify the impact of childhood sexual trauma on pelvic floor function. With a single pain with sex query, it is unclear what degree of pain or consistency of pain during sex would constitute a significant finding, leading to an affirmative response. Furthermore, epidemiological research questionnaires lack the detail of clinical queries.

Responses for each outcome varied over time in over half of the participants, with the minority of participants consistently responding affirmatively to urgency urinary incontinence (7%) and pain with intercourse (6%). We did not corroborate the reported symptoms with clinical records or testing. We did not inquire of other chronic pain and health problems associated with CSA. Although trauma patients are more likely to have additional pelvic floor complaints, including defecatory dysfunction, defacatory complaints were not assessed in SWAN.

While we conceptualized urgency urinary incontinence and pain with intimacy as high-tone PFDs and expected to find a causal relationship with CSA, only pain remained significant in our adjusted analyses. This finding may be due to the multiple contributors to urgency urinary incontinence beyond the pelvic floor like the detrusor muscle of the urinary bladder or a neurologic disorder, and the many possible symptomatic treatments like pelvic floor physical therapy or medications. BMI may play a role on the pathway from CSA to UUI outcome, but not pain outcome.

These results can help inform health provider's conversations with midlife women with one or more symptom of PFD. Specifically, women presenting with pelvic pain should be asked about childhood sexual trauma. Patients and providers may assume it is not necessary to discuss childhood events, and may not be comfortable discussing this sensitive topic in a brief clinical visit and examination. However, clinicians need to understand the importance of a complete medical history, regardless of age or race/ethnicity.

Such a history would include questions about ones' sexual abuse history and possible sequelae related to the pelvic floor, and can involve consultation with a behavioral health specialist who is knowledgeable about gynecological conditions. The physically focused biomedical treatment approach that relies heavily on expensive invasive testing of PFDs must be expanded to include psychosocial determinants to better understand and treat trauma-based PFD. Our data can inform the creation of a multidisciplinary team, including behavioral health providers to work toward the most optimal outcomes for patients.

Conclusions

This analysis of a prospective, multisite, multiethnic/multiracial cohort found that pain with sex, but not urinary incontinence was associated with childhood sexual trauma in multivariate analyses. Clinicians involved in assessments of PFD, such as pain with intercourse, should screen for childhood sexual trauma to facilitate an integrative approach. A multidisciplinary diagnostic and therapeutic approach is key and includes both psychological therapy and pelvic floor education. This approach can avoid unnecessary, expensive, and often times painful invasive tests, which may provoke re-traumatization.

Disclaimer

The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIA, NINR, ORWH, or the NIH.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The Study of Women's Health Across the Nation (SWAN) has grant support from the National Institutes of Health (NIH), DHHS, through the National Institute on Aging (NIA), the National Institute of Nursing Research (NINR), and the NIH Office of Research on Women's Health (ORWH) (Grants U01NR004061; U01AG012505, U01AG012535, U01AG012531, U01AG012539, U01AG012546, U01AG012553, U01AG012554, U01AG012495, and U19AG063720).