Uterine leiomyomata (UL) are a major contributor to gynecologic morbidity and medical costs, and black women are disproportionately affected by the condition. Previous studies have linked UL to psychosocial stress, including child abuse. We assessed the association between lifetime abuse victimization and UL among 9910 premenopausal women.
Data were derived from the Black Women’s Health Study, a prospective cohort study. In 2005, participants reported their experiences of physical and sexual abuse within each life stage (childhood, adolescence, adulthood). Biennial follow-up questionnaires from 2005 through 2011 ascertained new UL diagnoses. Rate ratios (RRs) and 95% confidence intervals (CIs) were estimated using Cox regression.
There were 1506 incident UL cases diagnosed by ultrasound or surgery. UL incidence was higher among women who reported child abuse, particularly sexual abuse. Relative to no abuse across the life span, RRs were 1.16 (95% CI, 1.02−1.33) for physical abuse only, 1.34 (95% CI, 1.09−1.66) for sexual abuse only, and 1.17 (95% CI, 0.99−1.39) for both physical and sexual abuse in childhood. RRs for 1-3 and 4 or more incidents of child sexual abuse were 1.29 (95% CI, 1.04−1.61) and 1.41 (95% CI, 1.07−1.85), respectively, whereas the RRs for low, intermediate, and high frequencies of child physical abuse were 1.19, 1.04, and 1.23, respectively. The association was strongest for the highest category of child abuse severity (RR, 1.57; 95% CI, 1.19−2.07). No associations were found for teen or adult abuse.
In the present study, child sexual abuse was an independent risk factor for UL, supporting the hypothesis that childhood adversity increases UL risk.
Uterine leiomyomata (UL) are clinically recognized in 30% of reproductive-aged women and are the primary indication for hysterectomy in the United States. Black women have 2-3 times the UL incidence of white women, have earlier ages at diagnosis, and have more symptomatic disease at clinical presentation. Differences in the prevalence of known risk factors do not explain the health disparity. Although UL etiology is poorly understood, sex steroid hormones are thought to influence their development and growth.
Emerging epidemiologic research suggests that exposure to psychosocial stress influences UL risk. In a cross-sectional study, a positive association was found between UL prevalence and both the number of major life events and stress intensity. Previously, we reported a positive association between perceived racial discrimination and UL in the Black Women’s Health Study (BWHS), a prospective cohort study. UL incidence was positively associated with greater levels of “everyday” racism and lifetime occurrences of major discrimination, with weaker associations found among those with higher coping skills. A positive association between early-life abuse victimization and UL was found in the Nurses’ Health Study II. Results were attenuated among those reporting high childhood emotional support, which like our study, supports the hypothesis that increased social/emotional support may buffer the negative impact of stress on health.
Severe stress in early life has been associated with altered regulation of the hypothalamic-pituitary-adrenal (HPA) axis and the autonomic stress response, which may persist into adulthood. Specifically, childhood sexual trauma and low social support have been linked to chronic HPA axis dysregulation. Stress can down-regulate the hypothalamic-pituitary support of ovarian hormones, but in some circumstances up-regulate adrenal progesterone. Studies of granulosa cell tissue culture also suggest that under certain conditions, ovarian steroid secretion increases rather than decreases. Various growth factors, cytokines, and matrix metalloproteinases involved in UL growth are up-regulated by stress hormones. Stress can also adversely affect health-related behaviors, leading to physical inactivity, heavy alcohol consumption, and poor diet, all of which can increase UL risk.
To assess whether there is an individual or cumulative stress burden of abuse across the life span, we prospectively evaluated the association of abuse victimization across 3 life stages (childhood, adolescence, and adulthood) with UL incidence in the BWHS. We assessed abuse according to timing, type, and frequency, and we also evaluated whether coping skills buffered these associations.
Materials and Methods
The Black Women’s Health Study is an ongoing prospective cohort study of 59,000 African American women aged 21-69 years at entry. In 1995, Essence magazine subscribers were mailed an invitation to enroll in a long-term health study by completing a comprehensive self-administered baseline questionnaire. Biennially, participants complete follow-up questionnaires to update their exposure and medical information; cohort retention has exceeded 80% through 2011. Participants reside in more than 17 states, with the majority residing in New York, California, Illinois, Michigan, Georgia, and New Jersey. The institutional review board of Boston University Medical Center approved the study protocol.
Assessment of abuse victimization
On the 2005 BWHS follow-up questionnaire, participants were asked questions about abuse victimization across the life span, including exposure as a child (up to age 11 years), teenager (aged 12-18 years), and adult (aged 19 years and older). The 9-item abuse instrument was adapted from the Conflict Tactics Scale and the Pregnancy Abuse Assessment Screen. Response categories were never, 1-3 times, or 4 or more times. We defined physical abuse as any report of a perpetrator having “pushed, grabbed, or shoved me,” “threw something at me that could hurt me,” “kicked, bit, or punched me,” “hit me with something including hand or fist,” or “physically attacked me in some other way” at a frequency of 4 or more times, or either “choked or burned me” or “seriously harmed someone I loved” at any frequency. We defined sexual abuse as any report of a perpetrator having “exposed genitals against my will” 4 or more times or “been sexual with me against my will” at any frequency.
We created a physical abuse summary score variable by assigning 1 point for each report of a physical abuse item occurring 4 or more times, with the exception of “choked or burned” or “seriously harmed someone I loved,” in which case 1 point was assigned for reports that these occurred 1-3 times and 2 points for reports that these occurred 4 or more times because these events were considered more severe. The resulting physical abuse severity score, which ranged from 0 to 9, was further categorized as low (score 1), intermediate (score 2), and high (score 3 or more).
We also created a summary variable for sexual abuse that classified 4 or more incidents separately from 1-3 incidents of sexual assault. Finally, within each life stage, we created a severity variable that defined mild abuse as 1 type of physical abuse that occurred 4 or more times; moderate abuse as sexual abuse 1-3 times and/or 2 forms of physical abuse occurring 4 or more times or a more severe form of physical abuse (“burn or choke or seriously harm someone I love”) occurring at any frequency; severe abuse as 3 or more types of physical abuse occurring 4 or more times and/or sexual abuse occurring 4 or more times; and very severe abuse includes 3 or more types of physical abuse occurring 4 or more times and sexual abuse occurring 4 or more times. These abuse definitions have been used in previous publications from the BWHS and other studies.
Assessment of uterine leiomyomata
On the 2005 questionnaire, women reported whether they had ever been diagnosed with uterine fibroids, the year of first diagnosis, and whether their diagnosis was confirmed by ultrasound and/or surgery (eg, hysterectomy), listed as 2 separate questions. On the 2007, 2009, and 2011 questionnaires, women reported whether their diagnosis had occurred in the previous 2 years. Cases were classified as surgically confirmed if they first reported confirmation by surgery, regardless of whether they also reported ultrasound confirmation. We also included cases confirmed by ultrasound because surgical cases represent only 10-30% of diagnosed cases. Ultrasound has a high sensitivity (99%) and specificity (91%) relative to the histologic evidence.
Assessment of covariates
On the 1995 and biennial follow-up questionnaires, we collected data on reproductive and contraceptive history, anthropometric factors, lifestyle factors (smoking, alcohol, physical activity), geographic region of residence, socioeconomic correlates (education, marital status, occupation), medical conditions, and Papanicolaou smear frequency. In 2003, women reported their household income. In 2007, participants reported on the recency of pelvic examination and pelvic ultrasound under 2 separate questions (never, less than 5, 5-9, or 10 or more years ago).
We also collected retrospective data on early life factors, including nativity (1997), maternal age (2005), sibship size and birth order (2007), parental education (2009), home ownership in childhood (2009), and passive smoke exposure (1997). In 2005, participants completed an abbreviated (10 item) scale based on an instrument developed by Carver designed to assess the following: (1) active coping, (2) use of emotional support, (3) use of instrumental support, (4) positive reframing, and (5) acceptance. Higher scores on the coping scale indicated higher coping skills.
We mailed supplemental questionnaires to a random sample of 248 cases regarding their initial date of diagnosis, method of confirmation, symptoms, and treatment and asked for permission to review their medical records. We obtained medical records from 127 of the 128 women who granted permission and corroborated the self-report in 122 (96%). Among the 188 cases with supplemental data (76%), 71% reported UL-related symptoms prior to diagnosis, and 87% reported that their condition came to clinical attention because they sought treatment for symptoms or a tumor was palpable during a routine pelvic examination. There were no appreciable differences between cases who did and did not release medical records with respect to UL risk factors.
We conducted a factor analysis of physical and sexual abuse items using an orthogonal rotation, which revealed 2 factors that confirmed the predetermined physical and sexual domains of abuse. Among 690 women who inadvertently returned a duplicate questionnaire in 2005, weighted kappa (κ) values indicated good reproducibility of responses to the abuse questions. In general, there was higher agreement in reporting of sexual abuse (child: κ = 0.68-0.78; teen: κ = 0.61-0.66; adult: κ = 0.51-0.59) than physical abuse (child: κ = 0.48-0.60; teen: κ = 0.47-0.59; adult: κ = 0.35-0.67), and child abuse was reported with greater reliability than teen or adult abuse. Cronbach alphas for the physical and sexual abuse items, respectively, were 0.80 and 0.82 in childhood, 0.81 and 0.72 in adolescence, and 0.84 and 0.63 in adulthood, indicating high internal consistency.
Of the 40,314 women who responded to both the 1997 and 2005 questionnaires, we excluded women who were postmenopausal (n = 14,824) or had been diagnosed with UL before 2005 (n = 12,907); women lost to follow-up after 2005 (n = 535); cases without a year of diagnosis (n = 116) or method of confirmation (n = 62); and women with incomplete data on abuse (n = 1534) or other covariates (n = 426), leaving 9910 women for analysis. Those excluded had lower education (mean: 15.7 vs 16.0 years) and were more likely to smoke (14% vs 10%) than those included but were similar with respect to age at menarche (mean: 12.4 vs 12.4 years), parity (67% vs 66%), age at first birth (mean: 22.7 vs 23.4 years), and other UL risk factors. The incidence rate for UL among the women who did not complete the abuse instrument was 27.7 cases per 1000 person-years, a rate similar to that among the nonabused women (28.1 cases per 1000 person-years).
We categorized abuse according to stage at first abuse, type, frequency, and severity (defined in previous text). The reference category for all analyses was “no abuse across the life span.” Incident cases were women who reported a first diagnosis of UL confirmed by ultrasound or surgery. Person-years at risk were calculated from the start of follow-up (March 2005) until UL diagnosis, menopause, death, loss to follow-up, or the end of follow-up (March 2011), whichever came first. Age- and period-stratified Cox regression models were used to estimate incidence rate ratios (RRs) and 95% confidence intervals (CIs) for the associations of interest.
We constructed 2 multivariable models. The first model controlled for age (1 year intervals) and time period (2 year intervals) in addition to early life factors, including nativity (foreign born, US born), maternal age (younger than 20, 20-24, 25-29, 30-34, 35 years or older), birth order (firstborn, later born), sibship size (children), childhood passive smoke exposure (yes, no), parental education (neither, at least 1 parent completed high school), home ownership in childhood (rented only, owned at some point), and age at menarche (years).
The second model further controlled for potential mediators, including parity (births), age at first birth (years), years since last birth (<5, 5-9, 10-14, 15, ≥20), oral contraceptive (OC) use (never, ever), age at first OC use (years), body mass index (BMI; <20, 20-24, 25-29, 30-34, ≥35 kg/m 2 ), smoking (current, past, never), current alcohol use (<1, 1-6, ≥7 drinks/week), education (≤12, 13-15, 16, ≥17 years), marital status (married/partnered, divorced/separated/widowed, single), occupation (white collar, non−white collar, unemployed), income (≤$25,000, $25,001-50,000, $50,001-100,000, >$100,000), region (South, Northeast, Midwest, and West), type 2 diabetes, and physical activity (metabolic equivalent [MET] hours per week). Factors that changed over time were modeled as time-varying covariates. Multivariable models mutually adjusted for physical and sexual abuse. Adolescent abuse was further adjusted for child abuse, and adult abuse was further adjusted for child and adolescent abuse. Further control for family history of UL or perceived racism made little difference in the effect estimates and were omitted from multivariable models.
We used stratification methods to examine whether associations were modified by other covariates such as coping, education, and income. P values from interaction tests were obtained using the likelihood ratio test comparing models with and without cross-product terms between abuse and covariates. Departures from proportional hazards were evaluated in the same manner using cross-product terms between abuse and age (<35, ≥35 years). Analyses were performed using SAS statistical software version 9.2 (SAS Institute, Cary, NC).
Baseline characteristics of participants according to type of abuse, within each life stage, are shown in Table 1 . Fifty-one percent reported abuse in childhood (43% physical and 19% sexual), 38% in adolescence (28% physical and 20% sexual), and 39% in adulthood (29% physical and 19% sexual). Relative to no abuse across the life span, child abuse was associated with younger age at baseline, younger maternal age, lower parental education, greater exposure to childhood passive smoking, lower prevalence of home ownership in childhood, earlier ages at first birth and first OC use, higher parity, lower adult household income, higher BMI, and greater tobacco and alcohol consumption. No appreciable differences were found for BMI at age 18 years or pelvic examination within the last 5 years. Patterns were generally similar for abuse reported in adolescence and adulthood.
|Stage and type of abuse|
|Characteristic a||No abuse across life span||Physical only||Sexual only||Physical and sexual||Physical only||Sexual only||Physical and sexual||Physical only||Sexual only||Physical and sexual|
|Number of women||2971||3161||629||1274||1802||1030||949||1961||934||953|
|Early life characteristics|
|Foreign born, %||5.7||6.0||6.6||6.5||5.9||6.9||5.5||5.3||6.4||5.6|
|First born, %||37.8||37.0||43.8||38.7||37.5||40.8||38.0||37.0||38.9||38.9|
|Sibship size, mean||3.9||4.1||3.9||4.3||4.2||3.9||4.4||4.2||3.9||4.2|
|Maternal age <20 y, %||12.7||16.2||18.5||24.3||18.9||17.5||24.8||18.9||14.4||20.1|
|Childhood passive smoke exposure, %||50.0||57.1||53.0||62.9||57.8||58.2||61.2||57.0||56.2||59.6|
|Neither parent finished high school, %||10.4||11.2||11.3||11.8||11.9||10.1||13.2||12.4||9.3||10.9|
|Home ownership in childhood, %||65.7||61.4||58.9||54.0||59.1||61.4||54.8||60.6||64.5||58.4|
|Age at menarche (y), mean||12.4||12.4||12.2||12.3||12.5||12.2||12.4||12.3||12.4||12.4|
|Age at baseline (y), mean||41.6||41.0||40.8||40.6||40.8||40.9||41.4||41.3||41.3||42.0|
|Age at first birth (y), mean b||26.1||25.5||25.0||23.8||24.6||24.5||23.0||23.8||25.7||23.5|
|Age at first OC use (y), mean c||19.7||19.4||19.4||19.1||19.4||19.1||18.6||19.1||19.6||18.7|
|Education (y), mean||15.0||15.1||15.0||14.9||14.9||15.0||14.7||14.7||15.2||14.7|
|Income ≤$25,000 in 2003, %||10||9||10||12||11||10||14||12||9||15|
|BMI at age 18 y (kg/m 2 ), mean||21.9||22.0||22.4||22.2||22.1||22.0||21.9||21.9||21.9||22.1|
|Current BMI (kg/m 2 ), mean||29.7||30.1||30.5||30.9||30.3||30.3||30.9||30.4||30.2||30.2|
|Physical activity (MET-h/wk), mean||20.8||20.8||20.9||21.1||21.8||19.5||21.7||20.7||20.8||21.5|
|Family history of UL, %||30||35||31||35||34||32||34||33||32||32|
|Smoking (current), %||8||11||9||11||14||13||19||16||13||18|
|Alcohol intake ≥7 drinks/wk, %||2||4||3||4||3||4||6||4||4||6|
|Coping score in 2005, mean||28.4||28.9||29.4||29.2||29.2||28.9||29.4||29.2||29.1||29.2|
|Pelvic exam <5 y in 2007, %||87||90||88||90||89||91||91||89||90||90|
During 47,382 person-years of follow-up, 1506 incident UL cases diagnosed by ultrasound (n = 1231) or surgery (n = 275) were reported. UL incidence was highest among women reporting child abuse, particularly sexual abuse ( Table 2 ). Relative to no abuse across the life span, RRs in the fully adjusted model were 1.16 (95% CI, 1.02−1.33) for physical abuse only, 1.34 (95% CI, 1.09−1.66) for sexual abuse only, and 1.17 (95% CI, 0.99−1.39) for both physical and sexual abuse in childhood. RRs increased monotonically with increasing frequency of child sexual abuse: RRs for 1-3 and 4 or more incidents of child sexual abuse were 1.29 (95% CI, 1.04−1.61) and 1.41 (95% CI, 1.07−1.85), respectively.
|Variable||Cases||Person-years||Age-adjusted RR (95% CI)||Multivariable RR (95% CI) a||Multivariable RR (95% CI) b|
|No abuse across the life span c||403||14,360||1.00 (referent)||1.00 (referent)||1.00 (referent)|
|Any abuse in childhood||842||24,030||1.22 (1.08–1.37)||1.20 (1.07–1.36)||1.19 (1.05–1.34)|
|Any abuse in adolescence||608||18,015||1.02 (0.84–1.24)||1.02 (0.83–1.24)||0.99 (0.81–1.21)|
|Any abuse in adulthood||600||18,146||1.03 (0.86–1.24)||1.03 (0.86–1.24)||1.01 (0.84–1.22)|
|Cumulative exposure to abuse|
|Childhood only||251||7112||1.22 (1.05–1.43)||1.22 (1.04–1.43)||1.22 (1.04–1.43)|
|Adolescence only||68||2329||1.03 (0.79–1.33)||1.04 (0.80–1.35)||1.02 (0.79–1.33)|
|Adulthood only||131||4454||1.06 (0.87–1.29)||1.06 (0.87–1.29)||1.05 (0.86–1.28)|
|Childhood and adolescence||184||5436||1.17 (0.98–1.39)||1.15 (0.97–1.37)||1.13 (0.95–1.35)|
|Childhood and adulthood||113||3441||1.16 (0.94–1.43)||1.15 (0.93–1.42)||1.13 (0.91–1.39)|
|Adolescence and adulthood||62||2210||1.01 (0.77–1.32)||1.01 (0.78–1.33)||1.00 (0.77–1.31)|
|All 3 life stages||294||8042||1.27 (1.09–1.47)||1.25 (1.08–1.46)||1.25 (1.07–1.46)|
|Stage at first abuse|
|Childhood||842||24,030||1.22 (1.08–1.37)||1.21 (1.07–1.36)||1.19 (1.06–1.35)|
|Adolescence||130||4538||1.02 (0.84–1.24)||1.03 (0.84–1.25)||1.01 (0.83–1.24)|
|Adulthood||131||4454||1.06 (0.87–1.29)||1.06 (0.87–1.29)||1.05 (0.86–1.28)|
|Type of abuse|
|Physical abuse only||511||15,111||1.18 (1.03–1.34)||1.17 (1.02–1.33)||1.16 (1.02–1.33)|
|Sexual abuse only||115||2892||1.38 (1.12–1.70)||1.38 (1.12–1.70)||1.34 (1.09–1.66)|
|Both physical and sexual abuse||216||6028||1.24 (1.05–1.46)||1.21 (1.02–1.43)||1.17 (0.99–1.39)|
|Physical abuse frequency b|
|Low||319||9309||1.19 (1.02–1.38)||1.19 (1.02–1.38)||1.19 (1.02–1.38)|
|Intermediate||143||4585||1.06 (0.87–1.29)||1.05 (0.86–1.28)||1.04 (0.85–1.27)|
|High||265||7245||1.24 (1.05–1.47)||1.23 (1.04–1.45)||1.23 (1.04–1.46)|
|Sexual abuse frequency e|
|1-3 incidents||216||5984||1.34 (1.08–1.66)||1.32 (1.07–1.65)||1.29 (1.04–1.61)|
|≥4 incidents||115||2936||1.43 (1.09–1.88)||1.41 (1.08–1.85)||1.41 (1.07–1.85)|
|Severity of abuse|
|Mild||248||7181||1.21 (1.04–1.42)||1.21 (1.03–1.42)||1.21 (1.03–1.42)|
|Moderate||275||7938||1.21 (1.03–1.40)||1.19 (1.02–1.39)||1.17 (1.00–1.36)|
|Severe||258||7642||1.16 (0.99–1.36)||1.14 (0.97–1.34)||1.13 (0.96–1.32)|
|Very severe||61||1270||1.66 (1.27–2.17)||1.60 (1.22–2.11)||1.57 (1.19–2.07)|
|Type of abuse|
|Physical abuse only||272||8709||0.95 (0.76–1.19)||0.95 (0.76–1.20)||0.92 (0.73–1.15)|
|Sexual abuse only||159||4908||1.03 (0.83–1.28)||1.03 (0.83–1.29)||1.03 (0.82–1.28)|
|Both physical and sexual abuse||177||4398||1.20 (0.93–1.54)||1.19 (0.92–1.53)||1.13 (0.87–1.47)|
|Physical abuse frequency d|
|Low||244||7159||0.96 (0.74–1.25)||0.97 (0.74–1.26)||0.95 (0.73–1.24)|
|Intermediate||88||2706||0.87 (0.62–1.22)||0.87 (0.62–1.22)||0.86 (0.61–1.21)|
|High||117||3242||0.94 (0.67–1.30)||0.94 (0.67–1.31)||0.95 (0.68–1.32)|
|Sexual abuse frequency e|
|1-3 incidents||280||7772||1.05 (0.82–1.34)||1.06 (0.83–1.35)||1.06 (0.82–1.35)|
|≥4 incidents||56||1534||1.00 (0.69–1.44)||1.00 (0.69–1.45)||0.96 (0.66–1.40)|
|Severity of abuse|
|Mild||157||5192||0.93 (0.73–1.18)||0.93 (0.73–1.18)||0.89 (0.70–1.13)|
|Moderate||304||8594||1.08 (0.87–1.33)||1.07 (0.87–1.32)||1.06 (0.86–1.31)|
|Severe||121||3683||0.95 (0.72–1.24)||0.94 (0.72–1.23)||0.91 (0.69–1.20)|
|Very severe||26||547||1.42 (0.91–2.21)||1.40 (0.90–2.18)||1.36 (0.86–2.12)|
|Type of abuse|
|Physical abuse only||299||9188||1.01 (0.83–1.24)||1.01 (0.83–1.23)||0.99 (0.81–1.21)|
|Sexual abuse only||147||4418||1.04 (0.83–1.30)||1.05 (0.84–1.31)||1.03 (0.82–1.29)|
|Both physical and sexual abuse||154||4541||1.06 (0.83–1.35)||1.07 (0.84–1.36)||1.08 (0.84–1.38)|
|Physical abuse frequency d|
|Low||276||8373||1.00 (0.81–1.25)||1.00 (0.80–1.24)||0.99 (0.80–1.24)|
|Intermediate||74||2306||0.95 (0.70–1.29)||0.96 (0.70–1.30)||0.97 (0.71–1.32)|
|High||103||3051||1.00 (0.74–1.33)||0.99 (0.74–1.33)||1.01 (0.75–1.36)|
|Sexual abuse frequency e|
|1-3 incidents||259||8021||0.97 (0.74–1.26)||0.98 (0.75–1.28)||0.96 (0.74–1.26)|
|≥4 incidents||42||938||1.33 (0.87–2.03)||1.39 (0.91–2.12)||1.38 (0.90–2.11)|
|Severity of abuse|
|Mild||211||6266||1.03 (0.84–1.25)||1.02 (0.84–1.25)||1.01 (0.83–1.22)|
|Moderate||288||9044||0.98 (0.81–1.18)||0.98 (0.81–1.19)||0.97 (0.80–1.18)|
|Severe||61||1732||1.06 (0.78–1.43)||1.06 (0.79–1.44)||1.07 (0.79–1.46)|
|Very severe||4||98||1.26 (0.46–3.41)||1.26 (0.46–3.43)||1.17 (0.43–3.21)|
a Adjusted for age, time period, foreign born, maternal age, birth order, sibship size, childhood passive smoke exposure, parental education, home ownership, and age at menarche and also adjusted for type and stage of abuse, when applicable;
b Further adjusted for parity, age at first birth, years since last birth, ever use of OCs, age at first OC use, BMI, current alcohol intake, smoking history, education, marital status, household income, occupation, type 2 diabetes, MET-hours per week of physical activity;
There was little evidence of a dose-response relation between frequency of child physical abuse and UL: relative to no abuse, RRs were 1.19, 1.04, and 1.23 for low, intermediate, and high frequency categories, respectively. The highest category of severity of child physical and sexual abuse was associated with a 57% increased rate of UL relative to no abuse (95% CI, 1.19−2.07). Abuse in adolescence or adulthood was not appreciably associated with UL incidence. Results did not vary significantly by age, education, income, or recent performance of pelvic examination or ultrasound (data not shown).
Associations between child sexual abuse frequency and UL were weaker among women with higher coping skills ( Table 3 ), defined as scoring above the median on the Carver coping scale. Among women with lower coping skills, RRs for 1-3 and 4 or more incidents of child sexual abuse were 1.50 and 1.98, respectively; among women with higher coping skills, the respective RRs were 1.19 and 1.05 ( P value for interaction = .08). Although stronger results were observed for child abuse overall, RRs were not statistically different across coping strata. Retrospective analyses that started follow-up in 1997 and included all incident cases from 1997 to 2011 ( Appendix ; Supplemental Table 1 ) or all prevalent and incident cases ( Supplemental Table 2 ) were weaker but still showed positive associations between child abuse and UL risk.
|Low coping skills (below median of Carver coping scale)||High coping skills (above median of Carver coping scale)|
|Variable||Cases||Person-years||Multivariable RR (95% CI) a||Cases||Person-years||Multivariable RR (95% CI) a||P value for interaction|
|No abuse across the life span b||171||6343||1.00 (referent)||208||7067||1.00 (referent)|
|Any abuse in childhood||357||10,173||1.23 (1.02–1.49)||458||12,968||1.14 (0.97–1.35)||.47|
|Type of abuse|
|Physical abuse only||214||6501||1.18 (0.96–1.45)||280||8045||1.13 (0.94–1.36)||.54|
|Sexual abuse only||52||1125||1.63 (1.18–2.24)||59||1631||1.16 (0.86–1.55)|
|Both physical and sexual||91||2547||1.20 (0.92–1.56)||119||3293||1.16 (0.92–1.47)|
|Physical abuse frequency c|
|Low||136||3977||1.21 (0.96–1.54)||172||4959||1.13 (0.91–1.40)||.94|
|Intermediate||58||1992||1.01 (0.74–1.38)||79||2420||1.06 (0.81–1.39)|
|High||111||3079||1.27 (0.98–1.66)||148||3959||1.20 (0.95–1.51)|
|Sexual abuse frequency d|
|1-3 incidents||90||2535||1.50 (1.07–2.12)||122||3242||1.19 (0.88–1.62)||.08|
|≥4 incidents||53||1137||1.98 (1.32–2.97)||56||1682||1.05 (0.70–1.57)|
|Severity of abuse|
|Mild||104||3083||1.23 (0.96–1.58)||136||3798||1.16 (0.91–1.49)||.79|
|Moderate||114||3355||1.18 (0.93–1.51)||152||4277||1.15 (0.90–1.46)|
|Severe||114||3255||1.22 (0.95–1.56)||136||4147||1.06 (0.81–1.39)|
|Very severe||25||481||1.79 (1.16–2.77)||34||747||1.44 (0.96–2.17)|